Regulating electron configuration of single Cu sites via unsaturated N,O-coordination for selective oxidation of benzene

Developing highly efficient catalyst for selective oxidation of benzene to phenol (SOBP) with low H2O2 consumption is highly desirable for practical application, but challenge remains. Herein, we report unique single-atom Cu1-N1O2 coordination-structure on N/C material (Cu-N1O2 SA/CN), prepared by water molecule-mediated pre-assembly-pyrolysis method, can efficiently boost SOBP reaction at a 2:1 of low H2O2/benzene molar ratio, showing 83.7% of high benzene conversion with 98.1% of phenol selectivity. The Cu1-N1O2 sites can provide a preponderant reaction pathway for SOBP reaction with less steps and lower energy barrier. As a result, it shows an unexpectedly higher turnover frequency (435 h−1) than that of Cu1-N2 (190 h−1), Cu1-N3 (90 h−1) and Cu nanoparticle (58 h−1) catalysts, respectively. This work provides a facile and efficient method for regulating the electron configuration of single-atom catalyst and generates a highly active and selective non-precious metal catalyst for industrial production of phenol through selective oxidation of benzene.

Reviewer #3 (Remarks to the Author): In this study, the authors synthesized a Cu single-atom supported on a N/C material with an unexpected coordination of Cu-N1O2. This catalyst exhibits a high selectivity for benzene oxidation to phenol, a challenging reaction for the chemical industry, with results close to the best catalysts in the literature (83.7% conversion and 98.1% selectivity). More impressive, phenol production is carried out with lower H2O2 amount (2:1 molar ratio) than other reports, which enhances the practical application of this material. The coordination of copper to oxygen atoms is possible due to the acidity of cyanuric acid in water, promoting the coordination via O-sites. The results seem to be relevant for the catalysis field, however, I am not convinced that the benzene conversion and products quantification were properly performed. In order to reach the required level to publish at Nature Communication, it is critical authors address question 5. Thus, the manuscript in the present form is not suitable for publication at Nature Communication. However, considering the relevance of the results, the manuscript might be after addressing the questions below: 1) "As shown in Fig. 1a, the melamine aqueous solution with copper nitrate was directly mixed with cyanuric acid aqueous solution, resulting in the Cu containing supermolecule precursor". The condensation of melamine and cyanuric acid occurs in aqueous media or in the pyrolysis step? What is the difference between light green powder before pyrolysis and the N/C material?
2) "vibration of aromatic ring located 1120 cm-1 of the Cu-N1O2 SA/CN precursor is weaker than that of CN precursor, indicating that Cu species is inlaid in the Cu-N1O2 SA/CN precursor27" The band at 1120 cm-1 is related exactly to which bond? Why the weakening of this band is related to Cu species? The given reference does not include this band in the discussion, not even in the SI. Moreover, it appears from Supplementary Fig.1 that this band does not exist in the Cu-N1O2.
3) "X-ray diffraction (XRD) patterns ( Supplementary Fig. 2) display the characterization diffraction peaks of graphic carbonitride (g-C3N4) at 13.0o and 27.4o with decreased diffraction intensity for Cu-N1O2 SA/CN contrast to CN and Cu NP/CN, implying the insertion of Cu species into CN matrix" The peak at 13° is too broad and weak to make any comparison, I would not even consider it as a peak. Also, the authors referred to the material as a carbon nitride, however, the N/C ratio (Supplementary Table 3) is too far from the ideal formula (C3N4) and probably the only peak observed (27°) is related to the stacking of N/C layers. The authors must name the material as a N/C material or N-doped carbon rather than a graphitic carbon nitride. 4) "The Cu content of Cu-N1O2 SA/CN was 0.16 wt%, determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES)." The copper loading is quite low, it seems that most of the inserted metal during the synthesis does not coordinate into the CN. What is the yield of copper addition for this synthesis?
5) The authors must provide the conversion equation for the benzene oxidation reaction as well as the other products found in the GC analysis. Does the conversion/selectivity include the possible CO2 formation? Is the carbon balance closed? The calibration curves used in the quantification must be presented in the SI. 6) In the Supplementary Fig.14, the authors showed that phenol oxidation is inhibit with Cu-N1O2 in comparison to Cu nanoparticles. However, the phenol oxidation conversion is not negligible, since around 10% is oxidized. What are the products found in the phenol oxidation reaction? How the consequent phenol oxidation does not influence the high selectivity of the reaction?
7) The benzene conversion under different reaction times must be provided (only 5h and 72h are shown) to understand the rate of phenol production. Nature Communication publications in the catalysis field are expected to present kinetic data.
8) The authors should carefully revise the text, some sentences in the manuscript contain typos and joined-words, i.e. "The good recycalbility is important for a heterogenious catalyst". Clearly, the text revision was not careful.

Response to the Reviewers' Comments
Reviewer #1 (Remarks to the Author): In this work, Zhao et al. prepared a series of Cu single-atom catalysts and studied the influence of the coordination environment of Cu sites on the performance in the oxidation of benzene. Differing in a N-coordination environment, the authors found that the single-atom Cu catalyst with Cu1-N1O2 coordination can efficiently boost the reaction at a low H2O2/benzene ratio of 2, and shows a higher turnover frequency than that of the Cu1-N2, Cu1-N3 and Cu nanoparticle catalysts. The conclusions obtained in the study are interesting; however, this work requires some significant improvement before publication is considered. Response: We appreciate the reviewer's comments. And all the comments and suggestions were considered carefully and incorporated completely in the revision (see below for a point-to-point response).  . Cu-N1O2 SA/CN features no Cu-Cu coordination (7.2 Å -1 ) of Cu foil, further demonstrating the isolated dispersion of Cu atoms. And Cu-N1O2 SA/CN displays only one intensity maximum at 4.6 Å -1 , which is between the Cu-O (4.9 Å -1 ) and Cu-N (4.2 Å -1 ) coordination in CuO and CuPc ( Supplementary Fig. 10). In contrast, the N coordinated Cu-N3 SA/CN and Cu-N2 SA/CN show intensity maximun at 4.3 and 4.4 Å -1 , respectively. Together with the concomitance of Cu-O and Cu-N bonds demonstrated by XPS results (Fig. 2b,c), the Cu atom was presumed to be coordinated by a mixed structure of Cu-O and Cu-N (Adv. Funct. Mater. 2021, 32, 2110224;Adv. Funct. Mater. 2022, 32, 2111446;ACS Catal. 2021, 11, 5212-5221). DFT calculations were further performed and gave the Cu1-N1O2 configuration as the most possible coordination structure (insert in Fig. 3d). The changes are highlighted in the revised version by giving the text a yellow background. The content of different N species in the as-prepared samples has been appended in Supplementary Table 11. As metallic single-atom was generally coordinated with pyridine N species [Adv. Funct. Mater. 2020, 30, 2000768;Chem. Soc. Rev., 2020, 49, 2215-2264Nat. Commun. 2020, 11, 5283], the pyridine N and Cu-N species were summed up as one. The results in Supplementary Table 11show slightly distinction among the as-prepared samples, however, no obvious relationship between this distinction and catalytic reactivity (Supplementary Table 7) was revealed. This result further demonstrates that the extrordinary catalytic performance of single-atom Cu-N1O2 SA/CN should be attributed to the unique Cu1-N1O2 sites. The changes are highlighted in the revised version by giving the text a yellow background.  Comment 4. At 1120 cm -1 of the FTIR spectrum, no aromatic ring absorption peak was found as the authors claimed. Response: Thanks for your comment. We accept the reviewer's comment that the peak at 1120 cm -1 was too weak to be observed even in enlarged view as shown in Fig.  R1.

Fig. R1
The enlarge view of FTIR spectra ofCN precursor and Cu-N1O2 SA/CN precursor.
For accurately describing the results of FTIR, the FTIR spectrum of Cu-N1O2 SA/CN catalyst was appended in the Supplementary Fig. 1, as shown followed.
Supplementary Fig. 1 The FTIR spectra of CN precursor, Cu-N1O2 SA/CN precursor and Cu-N1O2 SA/CN catalyst. Fourier Transform Infrared (FTIR) spectroscopy ( Supplementary Fig. 1) shows that the C=O stretching bands (νC=O) located at 1781 and 1741 cm -1 , which is higher than the reported νC=O of cyanuric acid (1739 and 1695 cm -1 ) [Chem.-Eur. J. 2009, 15, 6279-6288], indicating the formation of hydrogen-bonded supramolecular aggregates via hydrogen bonding of N-H…O and N-H…N linkages between melamine and cyanuric acid [Adv. Funct. Mater. 2013,23, 3661-3667]. Moreover, the Cu containing Cu-N1O2 SA/CN precursor shows similar FTIR spectrum to that of CN precursor, demonstrating that the presence of Cu 2+ does not affect the formation of hydrogen-bonded supramolecular aggregates. Followed by pyrolysis under N2 atmosphere at 600 o C for 2 h, the single-atom Cu-N1O2 SA/CN catalyst was obtained. FTIR spectrum of Cu-N1O2 SA/CN ( Supplementary Fig. 1) shows that the characteristic peaks of C=O disappear and new peaks centred at 3500-3000, 1800-1100, and 800 cm -1 attributed to tri-striazine arise, proving the presence of a heterocyclic ring structure [J. Am. Chem. Soc. 2003,125, 10288]. The changes are highlighted in the revised version by giving the text a yellow background.
Comment 5.The specific calculation methods for TON and TOF need to be provided. Response: Thanks for your comment. The specific calculation methods for TON and TOF have been appended in the Methods section and Supporting Information in the revised version and highlighted. Turnover frequency (TOF) of benzene was calculated as (mole of consumed benzene)/(reaction time (h)×mole of active Cu).
The turnover number (TON) of consumed H2O2 was calculated as (mole of consumed H2O2)/(mole of active Cu).
The mole of active Cu is determined by KSCN titration for single-atom Cu catalyst. For Cu NP/HCNS, the mole of active Cu is 50% for the 2 nm Cu nanoparticles.
The changes are highlighted in the revised version by giving the text a yellow background.  Supplementary Fig. 3 and Supplementary Table1"; In Line 222, "muchcloser" should be revised to "much closer". Response: Thanks for your constructive comment. The spelling and stylistic errors have been revised and highlighted in the revised version.

Reviewer #2 (Remarks to the Author):
The authors report a highly active and well-characterized SAC for selective benzene oxidation, with TOFs higher than 400 h -1 , competitive with benchmark catalysts. The catalytic results are remarkable and the theoretically-driven mechanistic insights offer an understanding of the catalytic steps. Unfortunately, the use of Cu SACs (on similar heterogeneous supports) for the hydroxylation of benzene into phenol submitted here is already reported in some other studies (some by the authors), both experimentally and theoretically, in line with this work, see  Chem. Soc. 2020, 142, 12643-12650. Therefore, given the less urgent and general nature of this nice piece of work, I suggest the publication of the communication in a specialized physical chemistry, catalysis or nanotechnology journal. Response: Thanks for your comment. In this manuscript, we present a highly active and well-characterized single-atom Cu catalyst for selective oxidation of benzene to phenol with high catalytic activity and phenol selectivity at 2:1 of a quite low molar ratio of H2O2/benzene. Moreover, we have in-depth studied the effect of coordination structure of single-atom site on its catalytic performance by theoretical and experimental methods, revealing the importance of electron modulation for designing high efficient catalyst.
Though the use of single-atom Cu catalsyts for benzene oxidation have been reported by some other studies, our work shows much process in both catalytic performance and kinetic understanding. Firstly, we realize highly-efficient benzene-tophenol transformation at a quite low H2O2 addition, which promotes the industrial production of phenol through selective oxidation of benzene. For another, this work also opens a new window for designing other single-atom catalysts with unique coordiantion structures towards diverse reactions.

Reviewer #3 (Remarks to the Author):
In this study, the authors synthesized a Cu single-atom supported on a N/C material with an unexpected coordination of Cu-N1O2. This catalyst exhibits a high selectivity for benzene oxidation to phenol, a challenging reaction for the chemical industry, with results close to the best catalysts in the literature (83.7% conversion and 98.1% selectivity). More impressive, phenol production is carried out with lower H2O2 amount (2:1 molar ratio) than other reports, which enhances the practical application of this material. The coordination of copper to oxygen atoms is possible due to the acidity of cyanuric acid in water, promoting the coordination via O-sites. The results seem to be relevant for the catalysis field, however, I am not convinced that the benzene conversion and products quantification were properly performed. In order to reach the required level to publish at Nature Communication, it is critical authors address question 5. Thus, the manuscript in the present form is notsuitable for publication at Nature Communication. However, considering the relevance of the results, the manuscript might be after addressing the questions below: Response: We appreciate the reviewer's constructive comments. And all the comments and suggestions were considered carefully and incorporated completely in the revision (see below for a point-to-point response). Fig. 1a, the melamine aqueous solution with copper nitrate was directly mixed with cyanuric acid aqueous solution, resulting in the Cu containing supermolecule precursor". The condensation of melamine and cyanuric acid occurs in aqueous media or in the pyrolysis step? What is the difference between light green powder before pyrolysis and the N/C material? Response: Thanks for your comment. In the synthesis procedure of Cu-N1O2 SA/CN, the melamine and cyanuric acid were condensed via hydrogen-bond interaction in aqueous media forming supermolecule precursor. With Cu atoms inlaid, it gave a light green color. The difference between the Cu-containing supermolecular precursor and Cu-N1O2 SA/CN catalyst was characterized by FTIR ( Supplementary Fig. 1).

Comment 1."As shown in
Supplementary Fig. 1 The FTIR spectra of CN precursor, Cu-N1O2 SA/CN precursor and Cu-N1O2 SA/CN catalyst.
Followed by pyrolysis under N2 atmosphere at 600 o C for 2 h, the single-atom Cu-N1O2 SA/CN catalyst was obtained. During thermal pyrolysis of the supermolecule precursor under protective gas, cyanuric acid reacted with ammonia to give ammelide, ammeline, and finally melamine. Meanwhile, intermediates like melem and melon was formed at 380 o C and 450 o C, respectively. The melon can further condense to yield carbon nitride in the temperature range of 450-600 °C. FTIR spectrum of Cu-N1O2 SA/CN shows that the characteristic peaks of C=O disappear and new peaks centred at 3500-3000, 1800-1100, and 800 cm -1 attributed to tri-s-triazine arise, proving the presence of a heterocyclic ring structure [J. Am. Chem. Soc.2003, 125, 10288].
The changes are highlighted in the revised version by giving the text a yellow background.
Comment 2."vibration of aromatic ring located 1120 cm -1 of the Cu-N1O2 SA/CN precursor is weaker than that of CN precursor, indicating that Cu species is inlaid in the Cu-N1O2 SA/CN precursor 27 " The band at 1120 cm -1 is related exactly to which bond? Why the weakening of this band is related to Cu species? The given reference does not include this band in the discussion, not even in the SI. Moreover, it appears from Supplementary Fig.1 that this band does not exist in the Cu-N1O2. Response: Thanks for your comment. We accept the reviewer's comment that the peak at 1120 cm -1 was too weak to be observed even in enlarged view as shown in Fig.  R1. For accurately describing the results of FTIR, the FTIR spectrum of Cu-N1O2 SA/CN catalyst was appended in the Supplementary Fig. 1, as shown followed.
Fourier Transform Infrared (FTIR) spectroscopy ( Supplementary Fig. 1) shows that the C=O stretching bands (νC=O) at 1781 and 1741 cm -1 , which is higher than the reported νC=O of cyanuric acid (1739 and 1695 cm -1 ) [Chem.-Eur. J. 2009,15, 6279-6288], indicating the formation of hydrogen-bonded supramolecular aggregates via hydrogen bonding of N-H…O and N-H…N linkages between melamine and cyanuric acid [Adv. Funct. Mater. 2013,23, 3661-3667]. Moreover, the Cu containing Cu-N1O2 SA/CN precursor shows similar FTIR spectrum to that of CN precursor, demonstrating that the presence of Cu 2+ does not affect the formation of hydrogen-bonded supramolecular aggregates. Followed by pyrolysis under N2 atmosphere at 600 o C for 2 h, the singleatom Cu-N1O2 SA/CN catalyst was obtained. FTIR spectrum of Cu-N1O2 SA/CN ( Supplementary Fig. 1) shows that the characteristic peaks of C=O disappear and new peaks centred at 3500-3000, 1800-1100, and 800 cm -1 attributed to tri-s-triazine arise, proving the presence of a heterocyclic ring structure [J. Am. Chem. Soc. 2003,125, 10288]. The changes are highlighted in the revised version by giving the text a yellow background.
Comment 3. "X-ray diffraction (XRD) patterns ( Supplementary Fig. 2) display the characterization diffraction peaks of graphic carbonitride (g-C3N4) at 13.0 o and 27.4 o with decreased diffraction intensity for Cu-N1O2 SA/CN contrast to CN and Cu NP/CN, implying the insertion of Cu species into CN matrix" The peak at 13° is too broad and weak to make any comparison, I would not even consider it as a peak. Also, the authors referred to the material as a carbon nitride, however, the N/C ratio (Supplementary Table 3) is too far from the ideal formula (C3N4) and probably the only peak observed (27°) is related to the stacking of N/C layers. The authors must name the material as a N/C material or N-doped carbon rather than a graphitic carbon nitride. Response: Thanks for your constructive comment. We accept the reviewer's comment that the characterization diffraction peak at 13 o in XRD pattern is too broad and weak to be considered as a peak. The peak observed at 27° is related to the stacking of N/C layers (Adv. Funct. Mater. 2013, 23, 3661-3667;Appl. Catal. B: Environ. 2023, 320, 121928). In the revised version, the matrix was named as N/C material (abbreviated to CN). The changes are highlighted in the revised version by giving the text a yellow background.
Comment 4. "The Cu content of Cu-N1O2 SA/CN was 0.16 wt%, determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES)." The copper loading is quite low, it seems that most of the inserted metal during the synthesis does not coordinate into the CN. What is the yield of copper addition for this synthesis? Response: Thanks for your comment. In the Cu-N1O2 SA/CN synthesis procedure, 40 mg of Cu(NO3)2·3H2O (containing 10.6 mg of Cu) was introduced. After pyrolysis treatment, about 0.15 g of Cu-N1O2 SA/CN with Cu loading of 0.16 wt% was obtained, representing 0.24 mg of Cu was inlaid in the final catalyst. Thus, the yield of copper addition in this synthesis is about 2.26%.
The low Cu loading might be due to the follows: In the supramolecular pre-assembly process of melamine with cyanuric acid, there is a competition between the hydrogenbond interaction of melamine with cyanuric acid and the Cu-O coordination of Cu ions with deprotonated cyanuric acid in alkaline aqueous solution. As a result, only a small number of Cu ions are coordinated and extracted from the Cu nitrate aqueous solution. Therefore, the Cu-N1O2 SA/CN has a low Cu content.
The changes are highlighted in the revised version by giving the text a yellow background.
Comment 5.The authors must provide the conversion equation for the benzene oxidation reaction as well as the other products found in the GC analysis. Does the conversion/selectivity include the possible CO2 formation? Is the carbon balance closed? The calibration curves used in the quantification must be presented in the SI. Response: Thanks for your constructive comment. The conversion equation for benzene oxidation has been appended in the Methods section and highlighted.
The conversion of benzene was calculated as (mole of consumed benzene)/(mole of initial benzene)×100%.
The selectivity of phenol was calculated as (mole of formed phenol)/(mole of consumed benzene)×100%.
The carbon balance was calculated as (mole of formed phenol + mole of formed benzoquinone + mole of remained benzene)/(mole of initial benzene)×100%.
The yield of phenol was calculated as (mole of formed phenol)/(mole of initial benzene)×100%.
The detail data was appended in Supplementary Table 7. As shown, based on our calculation equation, the carbon balance over single-atom catalysts were over 99%, indicating their weak deep oxidation ability for phenol. However, the nanoparticle Cu NP/CN displayed a low carbon balance of 85.4%, demonstrating its strong oxidation ability for organics degradation.
The changes are highlighted in the revised version by giving the text a yellow background. The calibration curves used in the quantification have been appended in the Supporting Information (See supplementary Fig. 13), as shown followed.