Oxo dicopper anchored on carbon nitride for selective oxidation of methane

Selective conversion of methane (CH4) into value-added chemicals represents a grand challenge for the efficient utilization of rising hydrocarbon sources. We report here dimeric copper centers supported on graphitic carbon nitride (denoted as Cu2@C3N4) as advanced catalysts for CH4 partial oxidation. The copper-dimer catalysts demonstrate high selectivity for partial oxidation of methane under both thermo- and photocatalytic reaction conditions, with hydrogen peroxide (H2O2) and oxygen (O2) being used as the oxidizer, respectively. In particular, the photocatalytic oxidation of CH4 with O2 achieves >10% conversion, and >98% selectivity toward methyl oxygenates and a mass-specific activity of 1399.3 mmol g Cu−1h−1. Mechanistic studies reveal that the high reactivity of Cu2@C3N4 can be ascribed to symphonic mechanisms among the bridging oxygen, the two copper sites and the semiconducting C3N4 substrate, which do not only facilitate the heterolytic scission of C-H bond, but also promotes H2O2 and O2 activation in thermo- and photocatalysis, respectively.

improve the manuscript: -First of all, the Introduction section can be improved in order to justify the selection of materials (Cu dimers and C3N4), in light of the hyphotesis posed in page 9. "We hypothesize that their non-integer oxidation state (intermediate between +1 and +2) and reduced cluster size (smaller Cu-Cu distance as compared to Cu-ZSM-5) would lend them exquisite catalytic performance for selective oxidation of methane".
-Concerning the comparison done in the thermo-catalytic approach with Cu-ZSM-5, although both have the dicopper-oxo centers, Cu oxidation states are different. Is there any other particular difference, such as Cu content, that could potentially have influence on the comparative results? -In the thermocatalytic approach, the authors demonstrate that gain factor, productivity and Cu content (of Cu oxocenters) are correlated, being sample with 0.35% the one with the highest oxygenated productivity. Has any additional attempt or preliminary test been carried out to increase Cu content and determine its effect on productivity?
-The proposed photocatalytic mechanism (Figure 4f) is not clear. Electron transfer from C3N4 to Cu dimer is suggested to take place, based on in situ XPS results, while holes are also illustrated to be transferred o the dimer, to oxidize CH4 leading to the formation of *CH3. Although all such processes would have specific dynamics, that mechanism would most likely lead to hole-electron recombination (comparable to type I heterojunctions). Please check, comment or clarify. Also, what do the values "-1.45 and 1.31 eV" represent?
Other minor comments: -Please indicate in Figure S1 what a) and b) parts correspond to. -At what temperature was the longer experiment in Figure S13b

Reviewer #1:
The authors have reported a new method to successfully synthesize oxo dicopper ([Cu-O-Cu] 2+ ) anchored on graphitic carbon nitride. The synthesized Cu2@C3N4 catalyst exhibits high selectivity for partial oxidation of methane under thermos-and photo-catalytic conditions. Many characterizations have been conducted to assure the structures of Cu2@C3N4 catalyst. It is an interesting work, I recommend the paper for publication after addressing the following issues. Response: We thank the reviewer for recommending publication of our work. We have followed the reviewer's suggestions to revise the manuscript and addressed the comments as detailed below.
Q1. There are lots of single atoms in the STEM images. Therefore, single atom Cu in C3N4 may affect the bond number analysis of EXAFS. How to exclude the single atoms contribution in the EXAFS characterizations. Response: We agree with the reviewer that we cannot fully exclude the presence of single-atom Cu in the Cu2@C3N4 catalysts but believe Cu dimers predominate in our case. Figure R1 below shows the comparison of EXAFS profiles for Cu1@C3N4 and Cu2@C3N4 (presented as a part of Figure S17 in the revised manuscript). Compared to Cu2@C3N4, Cu1@C3N4 lacks the second-shell Cu-Cu scattering feature at ~2.3 Å (prior to phase correction), while the primary Cu-N bonding (below 2 Å) feature is quite similar for the two catalysts. Motivated by the reviewer's comment, we have applied a linear combination of Cu monomers and dimers to fit the EXAFS spectrum of Cu2@C3N4 and derived a ratio of 72.4% for the Cu atoms in the dimeric configuration, versus 27.6% for monomers (Table 1, Figure 2d and Figure S6). This finding is largely in line with the statistical analysis based on STEM images, where the monomers and dimers were estimated to occupy ca. 46% and 54%, respectively, of the counts (corresponding to ~30% of Cu atoms in monomeric configuration). In the revised manuscript and supporting information, we have revised Table 1, Figure 2d, Figure S6 and the corresponding discussions accordingly. Figure R1. The Comparison of k 2 -weighted EXAFS spectra at Cu K edge for Cu1@C3N4 and Cu2@C3N4. Q2. How to exclude the single atoms contribution to the activity of partial oxidation of methane? Response: Considering that monomeric Cu in copper exchanged zeolites or supported on other substrates has also been reported to be active for CH4 oxidation, 1-3 we have conducted control experiments on single-atom Cu1@C3N4 ( Figure S17) for the catalytic studies. It was found that Cu1@C3N4 was merely active for methane oxidation, e.g., delivering a yield of only 0.03% (versus 0.2% by Cu2@C3N4) for methyl oxygenates at 50 o C (Figure 3b). The low reactivity of Cu monomers in partial oxidation of methane is further supported by the much lower gain factor of Cu1@C3N4 (0.03) than Cu2@C3N4 (0.19) for H2O2 utilization in partial oxidation of methane, which indicates H2O2 activation is more challenging on Cu monomers (Figure 3c and Figure S19). These results suggest that Cu monomers, if present in the Cu2@C3N4 catalysts, would not make significant contributions to the observed high methane partial oxidation activity, and also underlines the necessity of having dicopper-oxo centers for catalyzing the partial oxidation of methane under the reaction conditions investigated in this work.
Q3. Two hours catalyst test is not stable. Is the dicopper catalyst stable during the partial oxidation of methane reaction for a long time? Response: Motivated by the reviewer's comment, we have run durability tests with elongated periods. The methane conversions and product selectivities were implemented to Figure S13b of the revised manuscript. The methane conversion was found to increase monotonically with the reaction time. Although more CO2 was obtained at elongated reaction times, the selectivity toward methyl oxygenates was still as high as >82% after 6 h, indicating that the Cu2@C3N4 catalysts are stable under the reaction conditions (at 50 o C). The increase yield of CO2 can be attributed to the facile over-oxidation of methyl oxygenates as thermodynamically CO2 is the most stable product. Similar phenomena were also reported in the literature. 4,5 Notably, ICP analysis of the reaction solution after the 6-h durability test showed no Cu signal, ruling out the possibility of metal leaching during the reaction. Moreover, we have characterized the Cu2@C3N4 catalyst after the 6-h durability test by using aberration-corrected high-angle annular darkfield scanning transmission electron microscopy (HAADF-STEM) and X-ray absorption (XAS) spectroscopy. The collected STEM images still exhibited a lot of adjacent, paired bright dots that can be assigned to Cu dimers ( Figure S15a). EXAFS analysis for this spent Cu2@C3N4 catalyst, again with consideration of both monomeric and dimeric Cu sites, showed the coordination configuration was largely preserved (Table S2, in  comparison to Table 1). These results thus confirm the high stability of Cu2@C3N4 under the methane partial oxidation reaction conditions. Figure S13b. Time-dependent CH4 conversions (blue dots) and product selectivities (colored histograms) for the thermo-catalytic oxidation of CH4 with H2O2 at 50 o C. Table S2. Structural parameters being used to fit the EXAFS spectrum for spent Cu2@C3N4 after reaction with considering both of the contribution of single and dimeric Cu sites. Q4. How to calculate the formation energies in Table S1? More calculation details have to be given properly.

Response:
To calculate the formation energies of potential Cu dimer structures, we used the equation Here, E(System) represents the total free energy of the system, E(Substrate), E(Cu) and E(O2) represent the free energies of the support, Cu atoms and oxygen gas, and a and b are the numbers of Cu and O atoms involved in the considered structure. We have added such information to Table  S1 in the revised manuscript.
Q5. More DFT calculations, such as XPS and FTIR calculations, have to be performed to compare with experiment measurements to affirm the Cu2@C3N4 structure. Response: Following the review's suggestion, we calculated the Bader charge of Cu in Cu2@C3N4. It was found that the Cu atoms have oxidation states of +1.63 and +1.72 ( Figure S7), resembling the results derived from XPS and XANES spectra with an intermediate oxidation state between +1 and +2. The Cu dimers supported on C3N4 are hard to be detected by using FTIR, due to low loading of Cu (0.35 wt%) and the strong absorption of the C3N4 substrate. Instead, we have compared the experimental and calculated FTIR spectra for [Cu2(bpy)2(μ-ox)]Cl2, 2,2,-bipyridine and oxalic acid, in order to confirm the molecular transformation in the synthesis of Cu2@C3N4. The calculation was performed using the basis set of B3LYP/6-31g with Gaussian broadening. The simulated IR spectra ( Figure R2) are consistent with the experimental results ( Figure 1b). For instance, the asymmetric stretching vibration of C=C in the pyridyl ring of [Cu2(bpy)2(μ-ox)]Cl2 was predicted to give a FTIR peak at ~1630 cm -1 , close to the experimental result at 1650 cm -1 . Figure R2. The calculated IR Spectra of copper-dimer precursor [Cu2(bpy)2(μ-ox)]Cl2, 2,2,bipyridine and oxalic acid by gaussian simulation.
Q6. The meaning of energetic values in the reaction cycle in Figure 3 have to be illustrated clearly. It is better to present a potential energy surface diagram in the manuscript. Response: We thank the reviewer for pointing out this issue. The free energy diagram shown below was implemented as Figure 3e in the revised manuscript to illustrate the reaction pathway, with the energies given in Figure 3d representing the energy barriers associated with the given molecular transformations. For example, dissociations of the first (I-1 to I-2) and second (I-2 to I-3) H2O2 molecule has an energy barrier of 0.17 and 0.56 eV, respectively, corresponding to the transition states TS1 and TS2 in the predicted reaction pathway. Figure 3e. Predicted free energy diagram for the partial oxidation of CH4 with H2O2 on Cu2@C3N4. Three stages consisting of H2O2 activation, CH4 activation and methyl oxygenates formation are marked in different colors.

Reviewer #2:
This is an interesting paper. The topic of selective methane oxidation continues to attract a great deal of attention. This paper concerns a copper catalyst that uses hydrogen peroxide as the oxidant for thermal oxidation. This makes up teh bulk of the paper. There is a short section on using oxygen with photocatalysis. Response: We thank the reviewer for his/her interest in our work. We have followed the reviewer's suggestions to revise the manuscript and addressed the comments as detailed below. Q1. Cu is well known as an active site for methane oxidation. The novelty here is the use of a specially prepared Cu dimer catalyst. There is clear evidence that this is present in the fresh catalyst but there is no evidence presented as to whether this structure persists during or after the reaction. Does the complex break down and leach Cu into solution which would be a very good catalyst for making the reactive oxygen species OOH and OH radicals that they cite as the oxidising species. So they will need to provide evidence that the catalyst structure remains intact during teh reaction and that it is the sole species responsible for the observed activity. By teh way teh mechanism was originally proposed in reference 19 and that should be acknowledged. Response: Motivated by the reviewer' comment, we have performed ICP measurements on the spent Cu2@C3N4 catalyst (after 6 h of reaction, Figure S13b) and the solution after the reaction and confirmed no Cu leaching was detected. We have also characterized the spent catalysts using STEM and XAS. Similar to the situation prior to the catalytic studies, STEM images of the spent catalysts also exhibit paired bright dots, indicating the preservation of Cu dimers after the reaction ( Figure S15a in the revised manuscript). EXAFS analysis also showed that the atomic structure (including coordination number and bonding length) of Cu species in the spent catalysts (Figure S15b-S15d and Table S2) is consistent with the fresh ones ( Figure 2 and Table 1). These results thus confirmed the stability of Cu2@C3N4 under the reaction conditions, with the dioxo-copper centers being the active sites for partial oxidation of methane. In the revised manuscript, we have cited ref. 19 and acknowledged the originally proposed mechanism on page 13. Figure S13b. Time-dependent CH4 conversions (blue dots) and product selectivities (colored histograms) for the thermo-catalytic oxidation of CH4 with H2O2 at 50 o C. Table S2. Structural parameters being used to fit the EXAFS spectrum for spent Cu2@C3N4 after reaction with considering both of the contribution of single and dimeric Cu sites. Q2. The stability of hydrogen peroxide at 30 C and 70 C is questionable in the reaction times stated. Indeed, the "gain factor" is really low so what is the oxising species. If it is OH or OOH these radicals are very short lived so there are questions remaining on this aspect. Response: Although thermodynamically H2O2 is unstable and would spontaneously decompose into H2O and O2 at 30 or 70 o C (with △G of -152.5 and -153.7 kJ mol -1 , respectively), it is kinetically stable at these temperatures due to the relatively large dissociation barriers of ~0.6 eV on Cu dimers or ~1.2 eV on Cu monomers. The post-reaction titration of residual hydrogen peroxide using cerium sulfate showed that the conversion of H2O2 was as low as 6.2% at 30 o C and 12.4% at 70 o C, confirming its relatively high stability under the methane partial oxidation reaction conditions. We have performed additional control experiments at 50 o C using H2O or O2 as the oxidizing agent ( Figure R3). No oxidation products were detected in these cases, confirming H2O2 as the oxidizer for the partial oxidation of methane. We would like to point out that ‧OOH and ‧OH radicals are possible reaction intermediates derived from the dissociative activation of H2O2 on the Cu2@C3N4 catalysts, as detected using electron paramagnetic resonance (EPR) spectroscopy in the control experiments (with DMPO being used as the radical trapping agent, Figure S21) and confirmed by the reaction pathway predicted based on DFT calculations ( Figure 3). However, they are not necessarily the active species to attack CH4, which would otherwise cause kinetic hindrance due to the short lives of these radical species, as pointed out by the reviewer. Instead, in the predicted reaction pathway (Figure 3e), the Cu-dimer catalysts activate H2O2 and CH4 sequentially to form *OH (via ‧OH), *OOH (via ‧OOH) and *CH3, which then recombine to form methyl oxygenates. This pathway is not subjected to the possible kinetic limitation associated with the short-lived radicals. Figure R3. The comparison of catalytic activity for CH4 partial oxidation by using different oxidizing agents.

Reviewer #3 (Remarks to the Author):
This work reports on the preparation of dicopper oxocenters anchored on carbon nitride and its utilization in the selective oxidation of methane to oxygenated products, through thermo-and photo-catalytic approaches. The results indicate the formation of Cu dimers anchored on nitrogen groups, with particular structural properties and intermediate oxidation states. These materials demonstrate to reach relatively high methane conversion values under the evaluated conditions, by retaining high selectivity towards CH3OOH and CH3OH, over CO2, especially in the photocatalytic tests. Moreover, the authors properly correlate the intrinsic catalytic activation of H2O2 or O2, in the thermo-and photo-catalytic approaches, respectively. Certainly, the results are sound, and significant effort is done, in order to elucidate the correlation between activity and structure. However, some points indicated below could serve to clarify or improve the manuscript: Response: We thank the reviewer for recognizing our work. We have addressed the reviewer's comments in the revised manuscript, with the details provided below.
Q1. -First of all, the Introduction section can be improved in order to justify the selection of materials (Cu dimers and C3N4), in light of the hyphotesis posed in page 9. "We hypothesize that their non-integer oxidation state (intermediate between +1 and +2) and reduced cluster size (smaller Cu-Cu distance as compared to Cu-ZSM-5) would lend them exquisite catalytic performance for selective oxidation of methane". Response: We thank the reviewer for this constructive suggestion. We have revised the second paragraph of Introduction to rationalize the selection of C3N4 as substrate to support Cu dimers: "In this aspect, graphitic carbon nitride (g-C3N4) represents a promising photocatalytic substrate with a modest band gap in the range of 2.7-2.9 eV. [6][7][8] Its abundant nitrogen sites have been shown in many reports to be capable of anchor atomically dispersed transition metal sites. 9 It thus becomes interesting to investigate the potential coordination of active Cu sites on g-C3N4 and examine their synergies in partial oxidation of methane." Q2. Concerning the comparison done in the thermo-catalytic approach with Cu-ZSM-5, although both have the dicopper-oxo centers, Cu oxidation states are different. Is there any other particular difference, such as Cu content, that could potentially have influence on the comparative results? Response: We thank the reviewer for pointing out this question and agree that the dicopper-oxo centers in Cu2@C3N4 have multiple distinct atomic structures and electronic chemical properties from the counterpart in Cu-ZSM-5.
Besides the Cu sites exhibit different oxidation states (+1.7 vs +2), our combined EXAFS analysis and DFT calculations resolved the Cu-Cu distance to be ~2.71 Å (Table 1). It is noted that this value is much smaller than the Cu-Cu distance (~4.10 Å) associated with the dicopper-oxo center in Cu-ZSM-5. 10 In addition, compared to the case in Cu-ZSM-5, the Cu dimers supported on g-C3N4 have shorter Cu-O bond length (1.77 Å vs 1.88 Å) and smaller ∠Cu-O-Cu (99.6 o vs 135 o ). The distinct structure of Cu dimer in Cu2@C3N4 is believed to sterically favor the heterolytic cleavage of the C-H bond and facilitate the transfer of the -CH3 group for CH4 activation, evidenced by a lower energy barrier for the C-H bond dissociation (0.61 eV on Cu2@C3N4) than for Cu-ZSM-5 (~0.71 eV). 11,12 Moreover, the advantage of Cu2@C3N4 (vs Cu-ZSM-5) could also be attributed to the πconjugated heterocyclic rings and the semiconducting nature of the g-C3N4 substrate, which is known for accommodation of charge transfer between it and the anchored transition metal sites, which is beneficial for stabilization of the oxygenated adsorbates involved in the partial oxidation of methane with H2O2. Our DFT calculations show that the energy barriers for H2O2 activation to form ‧OOH and ‧OH are only 0.17 and 0.56 eV, respectively, substantially lower than the corresponding values (0.58 and 0.81 eV for ‧OOH and ‧OH, respectively) 4, 13 found for the dicopper-oxo centers confined in Cu-ZSM-5.
Q3. -In the thermocatalytic approach, the authors demonstrate that gain factor, productivity and Cu content (of Cu oxocenters) are correlated, being sample with 0.35% the one with the highest oxygenated productivity. Has any additional attempt or preliminary test been carried out to increase Cu content and determine its effect on productivity? Response: We indeed synthesized three Cu2@C3N4 catalysts with 0.25, 0.35 and 0.90 wt% of Cu to study the Cu loading effect on the productivity of methyl oxygenates ( Figure R4). The reactivity maximizes at 0.35 wt% Cu. Although we still cannot exclude the possibility of achieving higher catalytic activity at a higher (but lower than 0.90 wt%) loading, we believe adding more Cu species may cause cauterization of Cu and reduction of productivity. 14 Figure R4. The productivity of methyl oxygenates depended on the Cu loadings.
Q4. -The proposed photocatalytic mechanism (Figure 4f) is not clear. Electron transfer from C3N4 to Cu dimer is suggested to take place, based on in situ XPS results, while holes are also illustrated to be transferred to the dimer, to oxidize CH4 leading to the formation of *CH3. Although all such processes would have specific dynamics, that mechanism would most likely lead to hole-electron recombination (comparable to type I heterojunctions). Please check, comment or clarify. Also, what do the values "-1.45 and 1.31 eV" represent? Response: We thank the reviewer for raising this question. Charge transfer between the dimeric copper center and the C3N4 substrate was revealed by using in situ irradiation X-ray photoelectron spectroscopy (ISI-XPS). As shown in Figure 4d and Figure S27, shifts toward higher energies were observed at both O 1s and Cu 2p edges upon light irradiation, which was accompanied with a shift of the N 1s peak toward lower binding energies. Such phenomena consistently point to the transfer of holes (rather than electrons) from the g-C3N4 substrate to the dicopper-oxo center, where CH4 is activated and oxidized to form *CH3. Meanwhile, the excited electrons in the g-C3N4 substrate lead to the reduction of O2 to H2O2, which then migrates or diffuses onto the dicopperoxo center and gets activated to form *OOH or *OH. In the following, these oxygen species recombine with *CH3 to form methyl oxygenates, as in the case of thermocatalytic reactions. We have added the corresponding discussion to the revised manuscript (page 16).
The values "-1.45 and 1.31 eV" label the estimated position of dicopper-oxo states in the band structure of g-C3N4. This was determined by performing Tauc plot analysis on the UV-vis DRS and UPS spectra, 15,16 as shown in Figures S28 and S29. In the revised manuscript, we have revised the caption of Figure 4f to specify this.
Other minor comments: Q5. -Please indicate in Figure S1 what a) and b) parts correspond to. Response: Figure S1 exhibits the weight loss and the corresponding first derivative values of the dimeric copper complex (a) or the pristine Cu2@C3N4 with ligands (b) when burned in air during thermogravimetric analysis (TGA). We have revised the corresponding figure caption accordingly.
Q6. -At what temperature was the longer experiment in Figure S13b carried out? Response: The longer experiment was carried out at 50 o C. We have added the temperature info to the caption of Figure S13 in the revised manuscript.
Q7. -In last paragraph of page 11, please change Figure S2 by Figure S20. Response: We thank the reviewer for pointing out this error. We have corrected the corresponding figure citation.
Q8. -Please check: conversions are not shown in Figure 4a, but Yield (%), and the values at 1h are not included, nor selectivity. Response: We added the yield (at 1 h) in Figure 4a, and also put the CH4 conversions and selectivity of methyl oxygenates depending on time in the Supporting Information as Figure S25.
As stated in the review of previous version, this is a sound and interesting work dealing with the preparation of Cu-dimers anchored on carbon nitride structure, with significant activity towards methane oxidation to oxygenated products, through thermal and photocatalytic routes. In this revised version, the authors have addressed individual concerns and points related to the catalytic role, stability of dicopper centers and reaction mechanism. In this regard, the authors have conducted longer experiments proving the relative stability of these structures, and have provided more evidence of the proposed mechanism and intrinsic activities, including theoretical calculations. Also, a better explanation of the photocatalytic process has been added, with some experimental insights on the charge transfer from the semiconductor to the catalytic sites, assisted by thermal conditions. In general, this is a complete work with conclusive evidences of the favored role of catalytic centers towards scission of stable CH4 molecule and additional steps promoting the activation of the oxidizing species. Therefore, it not only highlights the intrinsic activity of the proposed structures, but also provides ideas for development of alternative materials able to promote such critical steps.