Universal Ligand-Assisted Supermolecular Strategy to Synthesize Single-Atom Catalysts for Solar-to-Ethanol Conversion

doi:10.21203/rs.3.rs-1665965/v1

PDF

Article

Universal Ligand-Assisted Supermolecular Strategy to Synthesize Single-Atom Catalysts for Solar-to-Ethanol Conversion

https://doi.org/10.21203/rs.3.rs-1665965/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted 24 May, 2022

You are reading this latest preprint version

Single-atom catalysts (SACs) have attracted significant attention; however, their catalytic properties are not satisfactory due to the sparse and inhomogeneity of the single-atom sites. Herein, we report a novel ligand-assisted supermolecular (LASM) strategy for the preparation of carbon nitride (CN)-supported SACs (M₁-CNs) with controllable and high active metal loadings. The ligands in the metal chelates can immobilise the metal and control their type and content, while the formed hydrogen-bonded supermolecules can inhibit the fusion of the CN precursor. Therefore, M₁-CN SACs (M: Cr, Mn, Fe, Co, Ni, Cu, Zn, Pd, Ag, and Cd) and multiple-metals SACs (Mα 1/Mβ 1-CNs: Fe/Cu and Cu/Zn/Ag) were successfully prepared. The as-synthesizd Cu₁-CN have a record-high loading of 9.75 at.%. Notably, Cu₁-CNs with high loadings achieved highly selective solar-driven photoreduction of CO₂ to ethanol, a breakthrough result in the field of 100% C₂₊ product construction using SACs, which stands among the best solar-to-ethanol single-atom photocatalysts up to date. Combining experimental evidence including X-ray absorption analysis and theoretical calculations, revealed that the Cu₁-CN with high loading promoted the C–C coupling of *CO intermediate products. This work provides a novel strategy for the controllable preparation of M₁-CNs and for the design of SACs with high selectivity and efficiency for the photoreduction of CO₂ to C₂₊ products.

As soon as the concept of single-atom catalysis was established¹, it became a topic of considerable research interest owing to the attractive prospect of achieving maximum atomic utilisation, high efficiency, high selectivity, and special catalytic mechanisms^2–4. Consequently, single-atom catalysis has been widely studied in the fields of electrocatalysis^5–7, photocatalysis^8,9, and thermal catalysis^10,11. In the past decade, significant efforts have been devoted toward single-atom catalysis^12–14, but the catalytic properties of single-atom catalysts (SACs) are far from those expected, particularly in terms of selectivity and multimolecular catalysis^15–18. For the selectivity of single-atom catalysis, in addition to metal single-atoms, the adjacent coordination environment of metal atoms is also an important factor because the coordination models have an important effect on the electronic structure of metal single-atoms and the adsorption of reactive molecules^19–22. However, the type^18,23, number^20,24,25, and state^26,27 of the adjacent coordination environment of metal atoms in the SACs are inhomogenous, leading to the poor selectivity of single-atom catalysis. For multimolecular reactions, it is possible to form new chemical bonds only when two molecules are close enough. However, the excessive space distance between active sites of the SACs, originating from the low single-atom loading, limit multimolecular reactions²⁸, resulting in the formation of simple and inexpensive products (e.g. H₂, CO) rather than complex and expensive ones (e.g. C₂H₄, C₂H₅OH) ^3,15,26. Hence, it is necessary to develop SACs with homogeneous and dense single-atom sites.

The number and uniformity of monatomic fixed sites on the support are crucial for SACs with a high loading and homogeneous active sites. Carbon nitride (CN) possesses an abundance of uniform monatomic fixed sites owing to a N content of 60 wt% and a definite heptazine structure²⁹, making it an ideal support for SACs with a high loading and high uniformity. When the prefabricated CN was used as the support of SACs, most of metal loadings were still less than 1 wt%^30,31. The CN formed in situ during the anchoring of metal atoms have showed high metal loadings. These CN-supported SACs with metal contents over 5 wt% had been successful preparated by pyrolysis of the mixture of metal salt and precursor of CN³². However, as the metal loading increased, metal clusters appeared in the CN-supported samples³³. Hence, most of the reported CN-supported SACs still suffer from poor selectivity and simple catalysis^34,35. The major reasons for this are the fusion of the precursor of CN and the migration of metal ions during thermal polymerisation under high temperatures, resulting in the inhomogenous and sparse active sites of CN-supported SACs³³. In theory, the fusion of the precursors of CN can be avoided by forming hydrogen-bonded supermolecules during thermal polymerisation because hydrogen-bond interactions can increase the precursors’ melting points³⁶. However, the metal ions cannot be controllably introduced into the supermolecules, especially in large quantities owing to the weak interaction between them. The ligand can fix metal ions by chelation³⁷ and form hydrogen bonds with CN’s precursors^8,38, thus achieving a controllable abundance of metal ions in the supermolecules. Therefore, we hypothesize that SACs with a high uniformity and high loading can be prepared using hydrogen-bonded supermolecules with the help of ligands.

Herein we developed a novel ligand-assisted supermolecule (LASM) strategy to prepare CN-supported SACs (M₁-CNs). The characterization results confirmed that this LASM strategy can be used for the large-scale synthesis of M₁-CNs containing high loadings of single (Cr, Mn, Fe, Co, Ni, Cu, Zn, Pd, Ag, or Cd) or multiple metals (Fe/Cu and Cu/Zn/Ag). This was achieved through metal–ligand complexations, infusible supermolecules, and the ultrahigh N content of the CN support. Notably, Cu₁-CN had a record-high loading of 9.75 at.% (or 34.3 wt%), which enabled the breakthrough photocatalytic reduction of CO₂ to pure ethanol. Theoretical calculations and experiments indicated that the ultrahigh Cu single-atom content promoted the C–C coupling reactions of *CO intermediate products over their hydrogenation.

Synthesis and characterization of M₁-CNs

The LASM strategy for the synthesis of M₁-CNs is shown in Fig. 1a. Briefly, the metal–ligand chelates were prepared by regulating the solution temperature, pH value, and types of metal salts and ligands. Then, a mixed solution of the metal–ligand chelates and cyanuric acid was added to a melamine solution to prepare the metal–containing supermolecule via hydrogen bond assembly among the metal–ligand chelates, melamine and cyanuric acid. After the pyrolysis of the supermolecule at a high temperature under an Ar flow, the M₁-CN was obtained³⁸. The X-ray diffraction (XRD) pattern and Fourier-transform infrared spectroscopy (FTIR) spectrum of the supermolecule (Supplementary Fig. 1) were different from those of the raw materials. Furthermore, the inerratic microrod morphology of the supermolecule was dissimilar from those of melamine and cyanuric acid (Supplementary Fig. 2), confirming the supermolecule formation^39,40. The colour of the sample and its XPS spectrum and elemental mapping show the homogeneous distribution of Cu in the Cu-containing supermolecule (Supplementary Fig. 3). In addition, the FTIR and XPS spectra demonstrated the presence of citric acid in the supermolecule (Supplementary Figs. 1b and 4). The roles of the supermolecule in the LASM strategy were then investigated by a series of experimental measurements. Thermogravimetry–differential scanning calorimetry (TG–DSC) analysis showed that the supermolecule has a higher decomposition temperature than melamine and cyanuric acid (Supplementary Fig. 5) because of its hydrogen bonds, as they prevent the fusion of supermolecule units under high temperatures and ensure the formation of CN (Supplementary Fig. 6a,b) (ref. 36). Hence, the metal atoms can be fixed during calcination to avoid their migration and aggregation. As a comparison, melamine easily melted during thermal polymerisation (Supplementary Fig. 6c,d), resulting in metal-atom agglomeration, even with a Cu content of 6.8 wt% (Supplementary Fig. 7). As a result, the samples obtained from the supermolecule displayed higher surface areas and pore volumes than the bulk CN that was derived from pure melamine (Supplementary Fig. 8 and Table 1). Moreover, the ligand plays a crucial role in ensuring the control and high loading of the metals in M₁-CNs. Additionally, the convenient operation and low cost of this method allowed for hundred-gram-scale syntheses of M₁-CNs (Supplementary Fig. 9), suggesting that the LASM strategy possesses great potential for industrial applications.

The XRD patterns and FTIR spectra confirmed that the support of Cu₁-CN is CN (Supplementary Fig. 10) (ref. 41). Except for the characteristic XRD peaks of CN, no other peaks were detected for Cu₁-CN, indicating the absence of crystals of Cu or its compounds⁴². Furthermore, Cu₁-CN demonstrates a nanosheet structure with a thickness of approximately 10 nm (Fig. 1b and Supplementary Fig. 11), with no observable Cu nanoparticles. The uniform dispersion of the Cu atoms in Cu₁-CN was verified by elemental mapping (Fig. 1c). The aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM) image clearly depicts high-density bright dots corresponding to Cu atoms without metal clusters and/or nanoparticles on the CN (Fig. 1d). The K-edge X-ray absorption near edge structure (XANES) spectrum of Cu in Cu₁-CN (Fig. 1e) revealed that the Cu oxidation state is between those of Cu₂O and CuO (Cu^δ+, 1 < δ < 2) (ref. 43,44). The charge was balanced by the loss of the H atom of CN (Supplementary Fig. 12). The Fourier-transformed k3-weighted X-ray absorption fine structure (XAFS) spectrum of Cu₁-CN (Fig. 1f) showed a solitary peak at 1.43 Å, which can be assigned to the Cu–N first coordination shell. Unlike that of Cu foil, the XAFS spectrum of Cu₁-CN did not exhibit a Cu–Cu coordination peak at 2.22 Å, indicating the absence of Cu clusters or nanoparticles in Cu₁-CN. From the wavelet transform (WT) contour plots (Supplementary Fig. 13), only the intensity maximum of the Cu–N bond at 4.18 Å⁻¹ was observed for Cu₁-CN. Notably, it was different from the Cu–Cu bond peak at 7.02 Å⁻¹ observed for the Cu foil. Cu within Cu₁-CN was determined to have a coordination number of approximately 3 and a bond length of 1.86 Å (Supplementary Table 2), demonstrating that the Cu atoms were each coordinated with three N atoms. These results indicate that Cu₁-CN was successfully prepared using the developed LASM strategy.

A series of SACs were then prepared and characterised to verify the versatility of the LASM strategy. We found that this strategy could precisely and continuously control the metal contents of the M₁-CNs; Cu₁-CN is used as an example for the following discussion. A series of Cu₁-CNs with different Cu contents were prepared by changing the copper salt and melamine molar ratios. XRD patterns and TEM images (Supplementary Fig. 14) revealed that none of the Cu₁-CNs contained Cu nanoparticles⁴². All the Cu atoms (bright dots highlighted by yellow circles in Supplementary Fig. 15) in the Cu₁-CNs showed excellent dispersion without clusters, even when the Cu content was as high as 9.75 at.% (or 34.3 wt%) (Fig. 1d), which is consistent with the XAFS results (Fig. 1f). Notably, this Cu loading (9.75 at.% or 34.3 wt%) is the highest metal loading reported thus far^45–47. Furthermore, the Cu contents of the Cu₁-CNs correlate well with the added dosage of copper salt (Fig. 1g), indicating that the developed strategy provides excellent control of the metal loading in SACs. When citric acid was not introduced during SAC preparation, the loading of Cu became very low and uncontrollable (Fig. 1g). Additionally, the UV-visible absorption peak of Cu²⁺ showed an obvious red shift when citric acid was added (Supplementary Fig. 16) but no changes occurred when cyanuric acid was added, proving that citric acid can coordinate with Cu²⁺ to form a metal–ligand chelate^37,48. Then, the metal–ligand chelate was introduced into the supermolecule via hydrogen bonding, which was confirmed by XPS, elemental mapping, and FTIR (Supplementary Figs. 1, 3, and 4). Overall, the ligand plays a key role in providing controllable, high metal loadings for the preparation of M₁-CNs. The developed LASM protocol can precisely and continuously control the metal loadings in M₁-CNs over a large range.

Next, the LASM strategy was extended to prepare M₁-CNs with different metals by changing the metal precursor species (e.g. Mn(Ac)₂, Co(NO₃)₃, or FeCl₃) and/or organic ligands (e.g. citric acid or oxalic acid); these were selected based on the stability of the resultant metal–ligand coordination complex (see the Supplementary Information for the detailed syntheses of M₁-CNs (M = Cr, Mn, Fe, Co, Ni, Zn, Pd, Ag, or Cd)) (ref. 37). The TEM images and XRD patterns of the as-prepared M₁-CNs (Supplementary Fig. 17) revealed that they possessed no metal nanoparticles or metallic crystals. The AC-HAADF-STEM images of all the M₁-CNs (Fig. 2) clearly indicated high-density bright dots representing the metal atoms, as well as an absence of metal clusters and/or nanoparticles. Furthermore, the X-ray absorption spectroscopy (XAS) results verified the monodispersity of metal atoms without M–M bonds in the M₁-CNs (Supplementary Fig. 18). These results demonstrate that the developed LASM strategy can serve as a universal protocol for SAC preparation.

This strategy could also be used to successfully synthesise multi-metal SACs. Specifically, dual-metal SAC Fe₁/Cu₁-CN and triple-metal SAC Cu₁/Zn₁/Ag₁-CN were prepared and characterised. The TEM images of Fe₁/Cu₁-CN (Fig. 3a and Supplementary Fig. 19) showed no metal nanoparticles or clusters; this result was supported by their XRD patterns. In addition, elemental mappings of Fe and Cu in Fe₁/Cu₁-CN exhibited the same distribution ranges as C and N (Fig. 3b), indicating that Fe and Cu were homodispersed in Fe₁/Cu₁-CN. The contents of Fe and Cu were 1.67 and 7.93 wt% (Supplementary Table 3), respectively. Bright dots representing Fe and Cu atoms were clearly observed in the AC-HAADF-STEM image of Fe₁/Cu₁-CN (Fig. 3c), suggesting that these atoms did not cluster and were atomically dispersed in Fe₁/Cu₁-CN. Furthermore, the XAFS spectrum of the as-prepared Fe₁/Cu₁-CN sample revealed the dispersions of its metal atoms (Fig. 3d,e). Specifically, no Fe–Fe or Cu–Cu bonds were observed, indicating that both Fe and Cu were atomically dispersed in Fe₁/Cu₁-CN. In addition, the XANES spectra revealed the oxidation states of Fe (Fe^δ+, 2 < δ < 3) and Cu (Cu^δ+, 0 < δ < 1), which were ascribed to the strong interactions of these metals with CN (Fig. 3f) (ref. 43). For Cu₁/Zn₁/Ag₁-CN, the TEM image and XRD pattern confirmed that no metal nanoparticles were present in the sample (Supplementary Fig. 20). According to the elemental mappings (Supplementary Fig. 21), the Cu, Zn, and Ag atoms had the same distributions as C and N, demonstrating that the three metals were uniformly distributed in Cu₁/Zn₁/Ag₁-CN. Furthermore, the results of inductively coupled plasma-atomic emission spectrometry (ICP-AES) confirmed the presence of Cu, Zn, and Ag in the Cu₁/Zn₁/Ag₁-CN sample (Supplementary Table 4). The AC-HAADF-STEM image of Cu₁/Zn₁/Ag₁-CN (Supplementary Fig. 22) revealed all the metals to be atomically dispersed, confirming the successfully preparation of this triple-metal SAC. The results obtained for Fe₁/Cu₁-CN and Cu₁/Zn₁/Ag₁-CN verify that multi-metal SACs can be prepared via the proposed LASM strategy.

These results demonstrate the capability of this novel LASM strategy to successfully synthesize SACs containing either single or multiple metals; this protocol also provides SACs with controllable, high metal loadings and can be implemented on a large scale. The four components greatly contributing to the versatility of this LASM SAC preparation strategy are summarised as follows. First, the metal–ligand complexations regulate the type of incorporated metal, which enables the preparation of SACs containing universal metals and multi-metal SACs. Second, the hydrogen bonds between the metal–organic chelate and melamine components allow for the metal contents in the supermolecule to be adjusted, which realizes controllable and high metal loadings. Third, the enhanced melting point of the raw material, originating from its hydrogen bonds, prevents the migration and agglomeration of metals. Fourth, CN produced in situ provides abundant single-atom fixed sites, which stabilise the single metal atoms. Unquestionably, the developed LASM strategy can provide SACs in addition to those prepared in this study.

Photocatalytic reduction of CO₂ to C₂H₅OH

The photoelectrical properties and catalytic activities of Cu₁-CNs were investigated. The single Cu atoms extended the light absorption range of CN and enhanced its absorption coefficient (Supplementary Fig. 23), providing Cu₁-CN with an improved solar energy utilisation efficiency. According to the photoluminescence (PL) spectra, Cu₁-CN has stronger fluorescence quenching and a shorter fluorescence lifetime than CN (Fig. 4a and Supplementary Fig. 24), indicating that Cu single atoms could remarkably reduce photogenerated charge recombination and accelerate charge separation and migration^17,49. As demonstrated by linear sweep voltammetry (LSV) (Fig. 4b), Cu₁-CN has a lower initial potential for CO₂ reduction than CN under a CO₂ atmosphere, indicating that Cu single atoms could reduce the energy barrier of CO₂ reduction⁵⁰. The current of Cu₁-CN is larger under a CO₂ atmosphere than under N₂, indicating that Cu₁-CN tends to reduce CO₂ rather than H⁺ (ref. 51,52). Furthermore, Cu₁-CN has a higher photocurrent than CN, allowing it to provide more photogenerated charges for redox reactions (Fig. 4c) (ref. 8,53). All these results indicate that Cu₁-CN has excellent potential for photocatalytic CO₂ reduction.

The photocatalytic products and yields of the as-prepared samples were tested. Owing to its poor absorption and high overpotential, CN showed no photocatalytic CO₂ reduction activity (Fig. 4d). However, as Cu is a common CO₂ reduction catalyst^54,55, the Cu₁-CNs did show photocatalytic CO₂ reduction activity. CH₃OH was the sole reduction product when the Cu single-atom content of Cu₁-CN was low. However, upon increasing the Cu content, C₂H₅OH was produced, indicating that high-density Cu sites play a crucial role in C–C coupling. When the Cu content reached 23.6 wt%, Cu₁-CN reduced the CO₂ to C₂H₅OH with 100% selectivity; notably, this is the first report on the catalytic reduction of CO₂ to pure C₂H₅OH using an SAC (Supplementary Table 5) (ref. 56,57). The ultrahigh selectivity of Cu₁-CN for C₂H₅OH is superior to those of most known nanocatalysts (Supplementary Table 6). At ultrahigh Cu contents (34.3 wt%), C₂H₅OH remains the sole reduction product, but its yield decreases because an excess of Cu atoms can reduce the generation of photo-induced charges and increase their recombination. The oxidative products were confirmed to be O₂ and tiny H₂O₂ (Fig. 4e and Supplementary Fig. 25), indicating that the photogenerated holes were consumed by oxidising water to produce H₂O₂, which then decomposed to O₂. No •OH was detected during the photocatalytic reaction (Supplementary Fig. 26), ruling out the adverse oxidation reaction from •OH. Namely, both the photogenerated electrons and holes can be used for the valuable reactions, indicating that Cu₁-CN possesses excellent application potential. Due to the solubility of O₂ in water (Supplementary Fig. 27), the ratio of O₂ to ethanol was lower than the stoichiometric ratio. To determine the source of C in C₂H₅OH, a ¹³C isotope labelling test was performed. When ¹³CO₂ was used, the molecular mass of C₂H₅OH was confirmed to be 48.1, indicating that the C in C₂H₅OH come from CO₂ rather than Cu₁-CN. The long-term stability of Cu₁-CN was evaluated by continuous CO₂ reduction for 20 h, and there was no obvious decay in either the ethanol or O₂ yield (Fig. 4e).

In situ attenuated total reflectance (ATR)-FTIR was conducted to characterize the reaction intermediates of the photocatalytic CO₂ reduction. For a scanning illumination time of 0–96 s, the Cu₁-CN spectra (Fig. 4f) displayed two bands at 1079 and 1355 cm^− 1, which were ascribed to the C–O stretching vibration of ethanol and the *COOH moiety of the intermediate product, respectively^58,59. During illumination, the intensities of the bands at 1580 and 1705 cm^− 1, which correlate to carbonate, continuously decreased⁵⁹; carbonate originates from the adsorption of CO₂ onto the alkalescent CN surface owing to residual amino groups. In the photocatalytic process, carbonate was rapidly consumed. When CN was used for CO₂ reduction, no intermediates were produced, even after 12 min of illumination (Supplementary Fig. 28). Based on these results, Cu plays an important role in the adsorption and reduction of CO₂, which is consistent with the photocatalytic activity analysis (Fig. 4d).

For heterocatalysis, adsorption between the reactants and catalysts is very important. As shown in Fig. 5a, there are no CO₂ temperature-programmed desorption (CO₂-TPD) signals for the CN sample, indicating that CN has a low CO₂ adsorption capacity. When the Cu single atoms were introduced, three strong CO₂-TPD signals were detected, and the CO₂ desorption temperature of the main peak was as high as 401°C, thus manifesting a strong adsorption energy between CO₂ and Cu₁-CN. Defect was a common CO₂ adsorption sites in reported works⁶⁰. However, the weaker electron paramagnetic resonance (EPR) signal of Cu₁-CN indicates that its defect density is lower than that of CN (Supplementary Fig. 29), which rules out the increase in CO₂ adsorption caused by an increase in the number of defects. Hence, CO₂ was well adsorbed at the Cu single-atom sites.

Finally, the adsorption energy of CO₂ relative to the adsorption site (Fig. 5b) was determined via density functional theory (DFT) calculations to illustrate the photocatalytic reduction of CO₂ to ethanol over Cu₁-CN. Among the Cu, C, and N sites, we found that the CO₂ adsorption energy is the lowest when it is adsorbed on the Cu single atom sites, indicating that CO₂ tends to adsorb on the Cu single atoms, in agreement with the experimental results (Fig. 5a). Owing to the weak adsorption capacity (Fig. 5b) and ultrahigh reduction overpotential (Fig. 4b) of CN to CO₂, CO₂ reduction cannot occur on the CN surface; thus, the CO₂ reduction process on CN was not considered. Because the Cu content of Cu₁-CN influences the products obtained via CO₂ reduction, two Cu₁-CNs models with low (approximately 13 wt%) and high (approximately 25 wt%) Cu contents (L-Cu₁-CN and H-Cu₁-CN, respectively) were constructed (Supplementary Fig. 30), and their reaction pathways were calculated (Fig. 5c). The intermediate product of *CO can only produce *CHO by hydrogenation on the L-Cu₁-CN rather than producing *OCCO through C–C coupling, owing to the overlarge distance between Cu active sites. Furthermore, L-Cu₁-CN showed a lower uphill energy barrier for the formation of *CHO than that of H-Cu₁-CN. For H-Cu₁-CN, the uphill energy barrier of *CO and *CO coupling to produce *OCCO (0.284 eV) was much lower than that of *CO hydrogenation to *CHO (1.718 eV). Thus, the ethanol product was obtained with H-Cu₁-CN, while methanol was obtained using L-Cu₁-CN. Importantly, it is these large energy barrier differences that provide the high selectivity of H-Cu₁-CN for producing ethanol.

A ligand-assisted supermolecule strategy was first developed to design and prepare CN-supported SACs. The employed ligands allow for the metal type to be controlled, and the supermolecules formed by hydrogen bonds prevent the fusion of the CN precursor. This strategy enables the controlled, large-scale preparation of SACs with high loadings of single metals (Cr, Mn, Fe, Co, Ni, Cu, Zn, Pd, Ag, or Cd) or multiple metals (dual- and triple-metal SACs). When Cu₁-CNs with high loadings were employed for photocatalytic CO₂ reduction, pure ethanol was produced with high selectivity, making this the first method to produce SACs that provide pure C₂₊ products via CO₂ reduction. DFT calculations show that the high-density Cu single atoms promote the C–C coupling of *CO intermediate products. This protocol provides a foundation for the design and preparation of novel SACs with high C₂₊ product selectivity, superior to that of previously reported nanocatalysts.

Chemicals. Copper (II) acetate monohydrate, melamine, oxalic acid and cyanuric acid were purchased from Aladdin Bio-Chem Technology Co., LTD (Shanghai). Iron chloride (III) hexahydrate was purchased from Shanghai McLean Biochemical Technology Co., Ltd. Citric acid was purchased from Tianjin Guangfu Technology Development Co., Ltd. CO₂ (99.999%) and Ar (99.99%) feed gases used in the experiment were purchased from Nanchang Guoteng Gas Co., Ltd. All reagents were used as received without further purification.

Preparation of M ₁ -CNs. Here, we take the synthesis of Cu₁-CN as an example. Melamine (1.00 g) and cyanuric acid (0.80 g) were dissolved in 100 and 120 mL deionized water by water bath heating, respectively. Citric acid (0.3362 g) and copper (II) acetate monohydrate (0.3193 g) were dissolved in 30 mL deionized water and then added into cyanuric acid solution for 5 min string. Then, this mixed solution was added into melamine solution. Followed 4 h stirring, precipitate was collected by centrifugation and dried at 60°C. Finally, the precipitate was heated to 600°C at a heating rate of 5°C·min^− 1 for 4 h in a tube furnace under argon flow. The obtained sample was denoted as Cu₁-CN. Furthermore, Cu₁-CNs with different Cu contents were prepared by adjusting the dosages of copper (II) acetate monohydrate (0.0399, 0.0799, 0.1597 or 0.3194 g). The preparative details of other M₁-CNs (M = Cr, Mn, Fe, Co, Ni, Zn, Pd, Ag, or Cd) were described in Supplementary Information.

Preparation of multi-metal SACs. Here, we take the synthesis of Fe₁/Cu₁-CN as an example. Melamine (1.00 g) and cyanuric acid (0.80 g) were dissolved in 100 and 120 mL deionized water by water bath heating, respectively. Oxalic acid (0.0720 g) and iron (Ⅲ) chloride hexahydrate (0.1081 g) were dissolved in 15 mL deionized water, and citric acid (0.1681 g) and copper (II) acetate monohydrate (0.0799 g) were dissolved in 15 mL deionized water. After that, these two solution were added into cyanuric acid solution for 5 min string. Then, this mixed solution was added into melamine solution. Followed 4 h stirring, precipitate was collected by centrifugation and dried at 60°C. Finally, the precipitate was heated to 600°C at a heating rate of 5°C·min^− 1 for 4 h in a tube furnace under argon flow. The obtained sample was denoted as Fe₁/Cu₁-CN. Furthermore, the preparative details of other multi metal SACs (Cu₁/Ag₁-CN and Cu₁/Zn₁/Ag₁-CN) were described in Supplementary Information.

Characterization. The crystalline phase of samples were measured by Bruker D8 advance X-ray diffraction (Cu Κα, λ = 1.541 Å) at a scanning speed of 2°·min^− 1. Fourier transform infrared (FTIR) spectra were recorded with a Vertex-70 spectrometer (Bruker). XPS measurements was carried out on a VG Escalab 250 (Thermo Fisher) instrument. SEM (XL30 S-FEG, FEI) and TEM (Tecnai F20) were used to examine the morphology of the as-prepared samples. TGA/SDTA851E thermogravimetric analyzer was used to carry out thermogravimetric analysis with a heating rate of 5°C·min^− 1 under argon flow. The specific surface areas and average porosities of the as-prepared samples were obtained by N₂ adsorption–desorption measurementson a NOVA 2000e (Quantachrome) equipment, using the Brunauer–Emmett–Teller (BET) method. JEM-ARM200F TEM/STEM with a spherical aberration corrector working at 300 kV recorded the images of HAADF-STEM and elements mapping. X-ray absorption spectroscopy experiment was carried out on BL08B2* of SPring-8 (8 GeV, 100 mA, Japan), where X-ray beam was monochromated by a water-cooled Si (111) double crystal monochromator and focused by two Rhcoated focusing mirrors with the beam size of 2.0 mm in the horizontal direction and 0.5 mm in the vertical direction around sample position. The samples of metals foil and corresponding metal oxides were used as reference, and all samples were measured by transmission mode. The spectra were analyzed and fitted using an analysis program Demeter. Wavelet simulation was carried out with continuous Cauchy wavelet transform method. The mass fraction of metals in the as-prepared catalysts were determined by an inductively coupled plasma atomic emission spectrometer (ICP-AES, ICAP 7200). UV–visible (UV–vis) spectra were recorded on a PE lambda 900 UV/VIS spectrophotometer equipped with an integrating sphere. Steady PL measurements were performed at 315 nm excitation on a Hitachi F-7000 spectrophotometer equipped with a Xenon lamp as excitation source. Time-resolved PL decay spectra were measured on FS5 fluorescence spectrometer (Edinburgh instrument) with a 380 nm picosecond pulsed diode laser (Epled-380) as excitation source. Mass spectrometry analysis were carried out on the GC-MS system (ISQ 7000) of Thermo Scientific single quadrupole GC-MS, equipped with mass filter, analyzer and direct injection rod, and triple off-axis Thermo Scientific* Dynamax* XR detection system. In-situ attenuated toal reflectance Fourier transform infrared spectroscopy (ATR-FTIR) was obtained by a Nicolet 6700 (Thermo Fisher) equipped a liquid nitrogen cooled MCT detector. Experiments were performed under the CO₂ and H₂O flow using a home-made ATR-FTIR setup. Temperature programmed chemisorption test was carried out on Autochem1 II 2920 to test the chemisorption of carbon dioxide. The defect concentration of the sample was measured by Brooke A300 electron paramagnetic resonance (EPR) spectrometer. The dissolved oxygen of reaction solution was measured by HACH HQ30d portable dissolved oxygen instrument.

Electrochemical measurements. LSV and photocurrent density plots were determined using a CHI-770 (Shanghai Chenhua) workstation with a three-electrode system in 0.5 M Na₂SO₄ electrolyte. The counter and reference electrodes were Pt plate and Ag/AgCl electrodes, respectively. The working electrode was prepared as follows: 5 mg catalysts were dispersed in mixed solution with 300 µL alcohol and 10 µL Nafion solution (0.5 wt%). Then, the solution was coated on a piece of fluorine-doped tin oxide glass (1×1 cm²) and then dried at 180°C for 12 h in a vacuum oven. Prior to and during all measurements, the electrolyte was purged with nitrogen.

Photocatalytic CO ₂ reduction test. Photocatalytic CO₂ reduction reaction was carried out in a closed glass system. The as-prepared Cu₁-CNs (5 mg) was dispersed into 100 mL pure water in the photoreactor and sealed. Before the light irradiation, CO₂ gas was continuously injected into the suspension at a rate of 100 mL·min^− 1 for 30 min to remove the air and provide a high concentration CO₂ gas in the reactor. Circulating water was used to maintain the reactor temperature at 6°C during photocatalytic reaction. Xenon lamp was used as light source, and 0.5 mL solution were taken every one hour. The products were detected by gas chromatography (GC7890A, Agilent Technology Co. Ltd.) equipped with flame ionization detector (FID) and thermal conductivity detector (TCD).

Data availability

The data that support the findings are available within the article and its Supplementary Information or from the corresponding authors on reasonable request.

Qiao, B., et al. Single-atom catalysis of CO oxidation using Pt₁/FeO_x. Nat. Chem. 3, 634–641 (2011).
Melike B., Adisak G., & Bruce C. G. Atomically dispersed metals on well-defined supports including zeolites and metal-organic frameworks: structure, bonding, reactivity, and catalysis. Chem. Rev. 120, 11956–11985 (2020).
Gu J., Hsu C., Bai L., Chen H., Hu X. Atomically dispersed Fe³⁺ sites catalyze efficient CO₂ electroreduction to CO. Science, 364, 1091–1094 (2019).
Zhang F., et al. Noble-metal single-atoms in thermocatalysis, electrocatalysis, and photocatalysis. Energy Environ. Sci. 14, 2954–3009 (2021).
Zhao Y., et al. Simultaneous oxidative and reductive reactions in one system by atomic design. Nat. Catal. 4, 134–143 (2021).
Zhang J., et al. Single platinum atoms immobilized on an MXene as an efficient catalyst for the hydrogen evolution reaction. Nat. Catal. 1, 985–992 (2018).
Lu Y., et al. Tuning the selective adsorption site of biomass on Co₃O₄ by Ir single atoms for electrosynthesis. Adv. Mater. 33, 2007056 (2021).
Jiang X., et al. Silver single atom in carbon nitride catalyst for highly efficient photocatalytic hydrogen evolution. Angew. Chem. Int. Ed. 59, 23112–23116 (2020).
Luo L., et al. Water enables mild oxidation of methane to methanol on gold single-atom catalysts. Nat. Commun. 12, 1218 (2021).
Li S., et al. Selective hydrogenation of 5-(hydroxymethyl) furfural to 5-methylfurfural over single atomic metals anchored on Nb₂O₅. Nat. Commun. 12, 584 (2021).
Ma Y., et al. High-density and thermally stable palladium single-atom catalysts for chemoselective hydrogenations. Angew. Chem. Int. Ed. 59, 21613–21619 (2020).
Ji S., et al. Chemical synthesis of single atomic site catalysts. Chem. Rev. 120, 11900–11955 (2020).
Shan J., et al. Short-range ordered iridium single atoms integrated into cobalt oxide spinel structure for highly efficient electrocatalytic water oxidation. J. Am. Chem. Soc. 143, 5201–5211 (2021).
Yao Y., et al. High temperature shockwave stabilized single atoms. Nat. Nanotechnol. 14, 851–857 (2019).
He Q., et al. Electrochemical conversion of CO₂ to syngas with controllable CO/H₂ ratios over Co and Ni single-atom catalysts. Angew. Chem. Int. Ed. 59, 3033–3037 (2020).
Wang J., et al. A single Cu-center containing enzyme-mimic enabling full photosynthesis under CO₂ reduction. ACS Nano 14, 8584–8593 (2020).
Li J., et al. Self-adaptive dual-metal-site pairs in metal-organic frameworks for selective CO₂ photoreduction to CH₄. Nat. Catal. 4, 719–729 (2021).
Li S., et al. Low-valence Zn^δ+ (0 < δ < 2) single-atom material as highly efficient electrocatalyst for CO₂ reduction. Angew. Chem. Int. Ed. 60, 22826–22832. (2021)
Jung E., et al. Atomic-level tuning of Co-N-C catalyst for high-performance electrochemical H₂O₂ production. Nat. Mater. 19, 436–442 (2020).
Yang T., et al. Coordination tailoring of Cu single sites on C₃N₄ realizes selective CO₂ hydrogenation at low temperature. Nat. Commun. 12, 6022 (2021).
Che H., et al. Iodide-induced fragmentation of polymerized hydrophilic carbon nitride for high-performance quasi-homogeneous photocatalytic H₂O₂ production. Angew. Chem. Int. Ed. 60, 25546–25550 (2021).
Zhang J., Xu X., Yang L., Cheng D., & Cao D. Single-atom Ru doping induced phase transition of MoS₂ and S vacancy for hydrogen evolution reaction. Small Methods 3, 1900653 (2019).
Sun X., et al. Phosphorus induced electron localization of single iron sites for boosted CO₂ electroreduction reaction. Angew. Chem. Int. Ed. 60, 23614–23618 (2021).
Fang S., et al. Uncovering near-free platinum single-atom dynamics during electrochemical hydrogen evolution reaction. Nat. Commun. 11, 1029 (2020).
Wang L., et al. Boosting activity and stability of metal single-atom catalysts via regulation of coordination number and local composition. J. Am. Chem. Soc. 143, 18854–18858 (2021).
Gong Y., et al. Regulating coordination environment of single-atom Ni electrocatalysts templated by MOF for boosting CO₂ reduction. Angew. Chem. Int. Ed. 58, 2705–2709 (2019).
Tiwari J., et al. Multicomponent electrocatalyst with ultralow Pt loading and high hydrogen evolution activity. Nat. Energy 3, 773–782 (2018).
Jin Z., et al. Understanding the inter-site distance effect in single-atom catalysts for oxygen electroreduction. Nat. Catal. 4, 615–622 (2021).
Wang Y., et al. Current understanding and challenges of solar driven hydrogen generation using polymeric photocatalysts. Nat, Energy 9, 746–760 (2019).
Huang P., et al. Selective CO₂ Reduction Catalyzed by Single Cobalt Sites on Carbon Nitride under Visible-light Irradiation. J. Am. Chem. Soc. 140, 16042–16047 (2018).
Li X., et al. Single-Atom Pt as Co-Catalyst for Enhanced Photocatalytic H₂ Evolution. Adv. Mater. 28, 2427–2431 (2016).
Lei G., et al.Isolated Iron Sites Embedded in Graphitic Carbon Nitride (g-C₃N₄) for Efficient Oxidative Desulfurization, Appl. Catal. B: Environ. 267, 118663 (2020).
An S., et al. High-density ultra-small clusters and single atom Fe sites embedded in graphitic carbon nitride (g-C₃N₄) for highly efficient catalytic advanced oxidation processes. ACS Nano 12, 9441–9450 (2018).
Wang Y., et al. Structure tuning of polymeric carbon nitride for solar energy conversion: from nano to molecular scale, Chem 5, 2775–2813 (2019).
Xu J., et al. Organic wastewater treatment by a single-atom catalyst and electrolytically produced H₂O₂. Nat. Sustain. 4, 233–241 (2021).
Han Q., et al. Atomically thin mesoporous nanomesh of graphitic C₃N₄ for high-efficiency photocatalytic hydrogen evolution. ACS Nano 10, 2745–2751 (2016).
Geoffrey A. Lawrance, Introduction to coordination chemistry, Wiley, 2010, ISBN: 978-0-470-51930-1.
Shalom, M., Inal, S., Fettkenhauer, C., Neher, D., Antonietti, M. Improving carbon nitride photocatalysis by supramolecular preorganization of monomers. J. Am. Chem. Soc. 135, 7118–7121 (2013).
Jun Y., et al. From melamine-cyanuric acid supramolecular aggregates to carbon nitride hollow spheres. Adv. Funct. Mater. 23, 3661–3667 (2013).
Zhang L., et al. Sodium-doped carbon nitride nanotubes for efficient visible light-driven hydrogen production. Nano Res. 11, 2295–2309 (2018).
Wang X., et al. A metal-free polymeric photocatalyst for hydrogen production from water under visible light. Nat. Mater. 8, 76–80 (2009).
Ma W., et al. Electrocatalytic reduction of CO₂ to ethylene and ethanol through hydrogen-assisted C-C coupling over fluorine-modified copper. Nat. Catal. 3, 478–487 (2020).
Kou Z., et al. Cage-confinement pyrolysis route to size-controlled molybdenum-based oxygen electrode catalysts: from isolated atoms to clusters and nanoparticles. Nano Energy 67, 104288 (2020).
Zhang L., et al. Carbon nitride supported high-loading Fe single-atom catalyst for activating of peroxymonosulfate to generate ¹O₂ with 100% selectivity. Angew. Chem. Int. Ed. 60, 21751–21755 (2021).
Wu J., Xiong, L., Zhao B., Liu M., Huang L. Densely populated single atom catalysts. Small Methods 3, 1900540 (2019).
Wu Q., et al. High-loaded single Cu atoms decorated on N-doped graphene for boosting Fenton-like catalysis under neutral pH. J. Mater. Chem. A 8, 13685–13693 (2020).
Li F., et al. Boosting oxygen reduction catalysis with abundant copper single atom active sites. Energy Environ. Sci. 11, 2263–2269 (2018).
Lu T., et al. Luminescent mechanochromic dinuclear Cu(I) complexes with macrocyclic diamine-tetracarbene ligands. Inorg. Chem. 57, 13618–13630 (2018).
Wei R., et al. Dual-cocatalysts decorated rimous CdS spheres advancing highly-efficient visible-light photocatalytic hydrogen production. Appl. Catal. B Environ. 231, 101–107 (2018).
Wang R., et al. Partial coordination-perturbed Bi-copper sites for selective electroreduction of CO₂ to hydrocarbons. Angew. Chem. Int. Ed. 60, 19982–19988 (2021).
Ouyang T., Ye Y., Wu C., Xiao K., Liu Z., Heterostructures composed of N-doped carbon nanotubes encapsulating Co and β-Mo₂C nanoparticles as bifunctional electrodes for water splitting. Angew. Chem. Int. Ed. 58, 4923–4928 (2019).
Zhang S., et al. Gradient hydrogen migration modulated with self-adapting S vacancy in copper-doped ZnIn₂S₄ nanosheet for photocatalytic hydrogen evolution. ACS Nano 9, 15238–15248 (2021).
Zhong Z., et al. Semi-chemical interaction between graphitic carbon nitride and Pt for boosting photocatalytic hydrogen evolution. Chin. Chem. Lett. 2021, Doi: 10.1016/j.cclet.2021.09.057.
Birdja Y., et al. Advances and challenges in understanding the electrocatalytic conversion of carbon dioxide to fuels. Nat. Energy 4, 732–745 (2019).
Chen S., et al. Lewis acid site-promoted single-atomic Cu catalyzes electrochemical CO₂ methanation. Nano Lett. 21, 7325–7331 (2021).
Wang G., et al. Photoinduction of Cu single atoms decorated on UiO-66-NH₂ for enhanced photocatalytic reduction of CO₂ to liquid fuels. J. Am. Chem. Soc. 142, 19339–19345 (2020).
Li J., et al. Self-adaptive dual-metal-site pairs in metal-organic frameworks for selective CO₂ photoreduction to CH₄. Nat. Catal. 4, 719–729 (2021).
Fu J., et al. Hierarchical porous O-doped g-C₃N₄ with enhanced photocatalytic CO₂ reduction activity. Small 13, 1603938 (2017).
Neatu S., Maciá-Agulló J. A., Concepción P., Garcia H. Gold-copper nanoalloys supported on TiO₂ as photocatalysts for CO₂ reduction by water. J. Am. Chem. Soc. 136, 15969–15976 (2014).
Rezaei M., Villalobos L., Hsu K., Agrawal K. Demonstrating and Unraveling a Controlled Nanometer-Scale Expansion of the Vacancy Defects in Graphene by CO₂. Angew. Chem. Int. Ed. 61, e 202200321 (2022).

Acknowledgements

We gratefully acknowledge the financial support of the National Natural Science Foundation of China (52100186, 52170082 and 52063024), the Natural Science Foundation of Jiangxi Province (20212ACB203008), Australian Research Council (ARC) through Future Fellowship (FT210100298), Discovery Project (DP220100603) and Linkage Project (LP210200504), and Japan Synchrotron Radiation Research Institute (2018A3631).

Author contributions

L.Z. T.M. and J.Z. conceived the ideas, designed the experimentals and co-wrote the manuscript. J.Z. and S.L. supervised the entire project. F.W., X.J., H.L. and S.Y. prepared, characterized and tested the catalysts. L.Z. and W.D. analysed the data. F.M. performed the AC-HAADF-STEM imaging. Q.S. and X.L. designed and performed the DFT calculations. M.W. and S.H. performed the in situ FTIR spectroscopy experiments and analysed the data. Y.C. performed the XAS data analysis. All the authors participated in discussion of the results and manuscript preparation and revision.

Competing interests

The authors declare no competing interests.

There is NO Competing Interest.

SupplementaryInformation.docx

PDF

Version 1

posted 24 May, 2022

You are reading this latest preprint version

Universal Ligand-Assisted Supermolecular Strategy to Synthesize Single-Atom Catalysts for Solar-to-Ethanol Conversion

Status:

Version 1

Abstract

Figures

Introduction

Results

Synthesis and characterization of M₁-CNs

Photocatalytic reduction of CO₂ to C₂H₅OH

Conclusion

Methods

References

Declarations

Competing Interests

Supplementary Files

Status:

Version 1

Universal Ligand-Assisted Supermolecular Strategy to Synthesize Single-Atom Catalysts for Solar-to-Ethanol Conversion

Status:

Version 1

Abstract

Figures

Introduction

Results

Synthesis and characterization of M1-CNs

Photocatalytic reduction of CO2 to C2H5OH

Conclusion

Methods

References

Declarations

Competing Interests

Supplementary Files

Status:

Version 1

Synthesis and characterization of M₁-CNs

Photocatalytic reduction of CO₂ to C₂H₅OH