Cooperative Dinuclear Activation of a Formate Intermediate in the Hydrogenation of CO₂ to Methanol

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript reports an interesting bimetallic catalytic system for CO2 hydrogenation to methanol under basic aqueous conditions. The concept is chemically compelling and potentially significant for the field. However, the current evidence does not yet fully support the proposed dinuclear mechanism, and several catalytic and mechanistic aspects require clarification. The manuscript requires following revision before it can be considered for publication.

1. The proposed Co-H/Ni-formate hydride-transfer pathway is the key novelty of the manuscript, but neither the cobalt hydride nor the dinuclear intermediate is directly observed. The stoichiometric reaction with Ni(OOCH)2 supports formate activation, but it does not conclusively prove a cooperative dinuclear mechanism. Additional in situ control experiments are needed.
2. Table 1 shows MeOH formation even in the absence of Ni, with MeOH TON = 1.6, and a small MeOH TON in the absence of Co. This conflicts with the statement that single catalysts produce only formate and no methanol. The authors should clarify whether these values are reproducible and above the detection limit.
3. All catalytic results are reported after 24 h. Time-dependent formate and methanol profiles are needed to confirm whether formate is truly an intermediate toward methanol or whether separate pathways are operating.
4. The formation of a black residue at higher Co loading suggests possible catalyst deactivation. The authors should provide evidence for catalyst stability and reproducibility, especially because the reported MeOH TONs are low.
5. Several presentation and reporting issues should be corrected before publication. Section numbering should be fixed, as Section 2.2 is missing; Scheme 5 should be clarified, particularly the identity, charge, and regeneration of the nickel species throughout the catalytic cycle. In addition, the statement that single catalysts produce “no methanol” should be revised because small MeOH TONs are reported in Table 1 and the SI. Finally, the reference list should be checked for formatting consistency and missing DOIs

Author Response

Reviewer 1

This manuscript reports an interesting bimetallic catalytic system for CO2 hydrogenation to methanol under basic aqueous conditions. The concept is chemically compelling and potentially significant for the field. However, the current evidence does not yet fully support the proposed dinuclear mechanism, and several catalytic and mechanistic aspects require clarification. The manuscript requires following revision before it can be considered for publication.

1. The proposed Co-H/Ni-formate hydride-transfer pathway is the key novelty of the manuscript, but neither the cobalt hydride nor the dinuclear intermediate is directly observed. The stoichiometric reaction with Ni(OOCH)2 supports formate activation, but it does not conclusively prove a cooperative dinuclear mechanism. Additional in situ control experiments are needed.

We have attempted several in situ control experiments, however, in our mechanistic studies we face two main limitations: (1) the paramagnetism of [Cp*Co(4DHBP)]I and Ni(acac)₂ complicating NMR studies and (2) the decomposition of [Cp*Co(4DHBP)]I at concentrations higher than (approximately) 2 mM under catalytic conditions.

Despite the paramagnetic character we explored NMR techniques for in situ experiments. When [Cp*Co(4DHBP)]I is pressurized in a NMR tube with 40 bar of H₂ in presence of D₂O and NaHCO₃, the only peaks detected are small peaks corresponding to the fraction of disassociated 4DHBP ligands.  No peak corresponding to the complex or the eventual Co-H are detected, despite the solution maintaining its distinctive purple color and forming no particulate. We propose the formation of a cobalt-hydride given the fact that [Cp*Co(4DHBP)]I is active towards formate production even without the presence of Ni(acac)₂ under basic conditions. We also base our hypothesis on the DFT calculations carried by Himeda et al. and Yang et al. (ref. 42 and 40 in the main text) on similar cobalt complexes.

The second technique explored for mechanistic studies is high-pressure FT-IR. While [Cp*Co(4DHBP)]I concentrations lower than 1 mM do not allow for signal detection, concentrations higher than 2 mM afford a dark precipitate.

1. Table 1 shows MeOH formation even in the absence of Ni, with MeOH TON = 1.6, and a small MeOH TON in the absence of Co. This conflicts with the statement that single catalysts produce only formate and no methanol. The authors should clarify whether these values are reproducible and above the detection limit.

The main text has been modified with the more correct statement of “no catalytic amounts of methanol”. The values are reproducible. The detection limit for methanol detection by GC-FID has been calculated to be 0.8 µmol (TON 0.27, using 3 µmol catalyst loading). The detection limit for formate has not been rigorously calculated but can be estimated around 1.5 µmol (TON 0.5, using 3 µmol catalyst loading).

1. All catalytic results are reported after 24 h. Time-dependent formate and methanol profiles are needed to confirm whether formate is truly an intermediate toward methanol or whether separate pathways are operating.

Additional experiments have been performed using different reaction time (t= 3, 6, 18, 24 hours). These results have been added to the SI (section II.C).

We find support for formate as intermediate in the following control experiments: 1) The catalytic tests reaction using [Cp*Co(4DHBP)]I alone, leads to the cobalt catalyzed hydrogenation of CO₂ to formate. 2) the stoichiometric hydrogenation of Ni(OOCH)₂ leads to the formation of methanol, supporting the claim that the nickel-formate is an intermediate towards methanol production. We clarified these points in the main text.

1. The formation of a black residue at higher Co loading suggests possible catalyst deactivation. The authors should provide evidence for catalyst stability and reproducibility, especially because the reported MeOH TONs are low.

The time-dependent profiles (Supporting Information, section II.C) show that catalysis is occurring throughout the 24 hour of reaction time, indicating that the catalysts are still active.

Additionally, from ESI-MS analysis of the reaction mixture only some ligand scrambling (Discussed in Section 2.3) occurs during the reaction.

1. Several presentation and reporting issues should be corrected before publication. Section numbering should be fixed, as Section 2.2 is missing; Scheme 5 should be clarified, particularly the identity, charge, and regeneration of the nickel species throughout the catalytic cycle. In addition, the statement that single catalysts produce “no methanol” should be revised because small MeOH TONs are reported in Table 1 and the SI. Finally, the reference list should be checked for formatting consistency and missing DOIs

These changes have been implemented in the main text and in the SI.

 

Reviewer 2 Report

Comments and Suggestions for Authors

In this manuscript, the authors focus on the challenging system of CO₂ hydrogenation to methanol under alkaline conditions, and proposes a Co–Ni bimetallic strategy to cooperatively activate the formate intermediate, aiming to overcome the key bottleneck of “stable formate species that are difficult to further hydrogenate.” Conceptually, the use of Ni to coordinate and activate formate while promoting Co–H transfer exhibits a certain degree of novelty. Preliminary mechanistic support is provided through comparative and stoichiometric experiments. However, based on the current level of data completeness and strength of evidence, the manuscript is not yet sufficient to substantiate its central mechanistic claims, and the conclusions largely remain at the level of phenomenological interpretation.

In summary, although the study presents a conceptually appealing strategy for CO₂ hydrogenation, the current manuscript lacks mechanistic rigor, depth of characterization, and sufficient experimental validation of the proposed reaction pathway. In its present form, the work does not meet the evidentiary standard required to support its central mechanistic conclusions, and substantial additional mechanistic and catalytic evidence is necessary for publication consideration.

1. The reaction is conducted at 50 bar (CO₂/H₂) in a mixed H₂O/THF solvent system. Under the current conditions, could mass transport limitations, rather than intrinsic catalytic control, be involved?
2. It is well known that THF may undergo ring-opening or oxidative degradation under such conditions. Have the authors experimentally verified the chemical integrity of THF during reaction? Is there any evidence that THF-derived fragments do not participate in or interfere with the carbon balance or reaction pathway?
3. The manuscript does not provide direct evidence clarifying how H₂ is activated within the catalytic system. Whether hydrogen dissociation occurs via homolytic cleavage on metallic sites or heterolytic activation involving interfacial species remains unclear.
4. Under H₂ and alkaline conditions, is it possible that the actual catalytically active species is Ni⁰ rather than the proposed Ni²⁺ coordination complex?
5. The proposed methanol formation pathway is primarily constructed based on literature precedent rather than direct experimental validation. For the methanol formation process, additional systematic control experiments are still required to substantiate and strengthen the proposed mechanistic conclusions.6.
6. some useful literatures should be refered to further investigate the mechasim from other characterizations, like Nano Research,2024, 17 (8), 7194-7202; Advanced Materials, 2024, 44, 2410125.

Author Response

We thank the reviewer for taking the time to evaluate our manuscript.

 

Reviewer 2

In this manuscript, the authors focus on the challenging system of CO₂ hydrogenation to methanol under alkaline conditions, and proposes a Co–Ni bimetallic strategy to cooperatively activate the formate intermediate, aiming to overcome the key bottleneck of “stable formate species that are difficult to further hydrogenate.” Conceptually, the use of Ni to coordinate and activate formate while promoting Co–H transfer exhibits a certain degree of novelty. Preliminary mechanistic support is provided through comparative and stoichiometric experiments. However, based on the current level of data completeness and strength of evidence, the manuscript is not yet sufficient to substantiate its central mechanistic claims, and the conclusions largely remain at the level of phenomenological interpretation.

In summary, although the study presents a conceptually appealing strategy for CO₂ hydrogenation, the current manuscript lacks mechanistic rigor, depth of characterization, and sufficient experimental validation of the proposed reaction pathway. In its present form, the work does not meet the evidentiary standard required to support its central mechanistic conclusions, and substantial additional mechanistic and catalytic evidence is necessary for publication consideration.

1. The reaction is conducted at 50 bar (CO₂/H₂) in a mixed H₂O/THF solvent system. Under the current conditions, could mass transport limitations, rather than intrinsic catalytic control, be involved?

As the rate and the TONs are low, there is no significant pressure drop during the reaction and the consumption of the gasses is slow compared to the mixing, so mass transport limitations are not playing a role under these conditions.

1. It is well known that THF may undergo ring-opening or oxidative degradation under such conditions. Have the authors experimentally verified the chemical integrity of THF during reaction? Is there any evidence that THF-derived fragments do not participate in or interfere with the carbon balance or reaction pathway?

Blank experiments involving THF and NaHCO₃ in THF did not show any product or degradation when analyzed by GC and ¹H NMR. These have been now added to the Supporting Information (Table S1, entries 10, 11). THF has been reported in various systems applied for the hydrogenation of CO₂ to methanol or formate, with no account to the best of our knowledge commenting on its degradation under comparable conditions. Examples may be found in references: 27, 29-31, 33, 34, 38.

1. The manuscript does not provide direct evidence clarifying how H₂ is activated within the catalytic system. Whether hydrogen dissociation occurs via homolytic cleavage on metallic sites or heterolytic activation involving interfacial species remains unclear.

Due to the paramagnetism of [Cp*Co(4DHBP)]I and its decomposition at higher concentrations, it is not possible for us to detect the proposed Co-H species (see also comments ref 1). Based on reported literature DFT studies (references 40 and 42), we propose a heterolytic splitting of H₂ to form a cobalt-hydride and a protonated OH on the bipyridinol ligand. It was observed that when using less base the activity of [Cp*Co(4DHBP)]I in the hydrogenation of CO₂ to formate significantly decreased (Table S1, entry 7), in line with our hypothesis that the protonation of the bipyridinol ligand is necessary in the catalytic cycle.

Formation of a Co-(H₂) bond and homolytic splitting to form Co(-H)₂ would not be possible due to insufficient free coordination sites. A similar cobalt complex, [Cp*CoI(5,5'-dimethyl-2,2'bipyridyl)]I does not present this protonation-deprotonation mechanism and would form the required cobalt hydride via homolytic splitting. When [Cp*CoI(5,5'-dimethyl-2,2'bipyridyl)]I was tested, both formation of formate and methanol did not occur (Table S1, entry 9). No activity was observed in the presence of Ni(acac)₂ (Supporting Information, Catalytic activity of [Cp*CoI(5,5'-dimethyl-2,2'bipyridyl)]I).  

Altogether we believe this is collectively is sufficient support to propose the mechanism. Our formulations in the paper are rather careful.

1. Under H₂ and alkaline conditions, is it possible that the actual catalytically active species is Ni⁰ rather than the proposed Ni²⁺ coordination complex?

As shown in Table 1, entry 7, catalytic activity towards methanol and formate production has been detected using Ni(COD)₂ (albeit low). We can conclude that Ni(acac)₂ and Ni(COD)₂ are probably accessing the same active species in the reaction mechanism. Therefore, we cannot exclude that Ni⁰ is the catalytically active species. We have emphasized this in the article.

1. The proposed methanol formation pathway is primarily constructed based on literature precedent rather than direct experimental validation. For the methanol formation process, additional systematic control experiments are still required to substantiate and strengthen the proposed mechanistic conclusions.6.

The hydrogenation of formaldehyde to methanol has been studied and has been observed to occur by both [Cp*Co(4DHBP)]I and Ni(acac)₂ both under single metal conditions and in combination (Supporting Information, Table S4) in higher amounts than the regular decomposition of formaldehyde under the used reaction conditions.

1. some useful literatures should be refered to further investigate the mechasim from other characterizations, like Nano Research,2024, 17 (8), 7194-7202; Advanced Materials, 2024, 44, 2410125.

In situ IR measurements have been attempted for this system, but failed due to the decomposition of [Cp*Co(4DHBP)]I at the concentrations required for signal detection. The other techniques described in the aforementioned reports do not apply to liquid homogeneous systems under pressures higher than 1 bar.

 

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript presents a novel approach to activating formate under basic conditions using the bimetallic (Co, Ni) system, allowing the production of methanol without using excessive amounts of additives. The concept of using Ni to activate formate and facilitate hydride transfer from Co is novel and is well supported by the experiments presented.

There are some issues that the authors should address before publication in the Molecular:

1. The TONs reported by the authors must be compared with the literature
2. What about the selectivity to methanol under the studied conditions?
3. Did the authors study the effect of pH on activation?
4. The proposed bimetallic mechanism requires further validation, as it lacks direct spectroscopic evidence for key intermediates, specifically the active Co–H species and the activated Ni–formate complex. To substantiate the claimed bimetallic hydride transfer, the authors should incorporate in situ techniques such as IR, paramagnetic NMR, or provide a detailed justification for the absence of this data.

Author Response

We thank al reviewers for taking the time to evaluate the work described in this manuscript.

 

Reviewer 3

The manuscript presents a novel approach to activating formate under basic conditions using the bimetallic (Co, Ni) system, allowing the production of methanol without using excessive amounts of additives. The concept of using Ni to activate formate and facilitate hydride transfer from Co is novel and is well supported by the experiments presented.

There are some issues that the authors should address before publication in the Molecular:

1. The TONs reported by the authors must be compared with the literature

These changes have been implemented in the main text.

1. What about the selectivity to methanol under the studied conditions?

Formate salt and methanol were the only two products detected after the catalytic reactions.

1. Did the authors study the effect of pH on activation?

The pH of the reaction mixture was not measured as these are not aqueous solutions, but we have varied the concentration of base. Under standard conditions, the pH is estimated to be 8.6. When using less base, the activity of [Cp*Co(4DHBP)]I in the hydrogenation of CO₂ to formate significantly decreases both alone and in combination with Ni(acac)₂ (Table S1, entry 7 and Table 1, entry 8, respectively). Meanwhile, increasing the amount of base is limited by its solubility in the solvent mixture. When increasing the base to 8 mmol, the remaining solid NaHCO₃ hindered the stirring of the reaction, therefore decreasing the formate TON (Table S1, entry 6).

1. The proposed bimetallic mechanism requires further validation, as it lacks direct spectroscopic evidence for key intermediates, specifically the active Co–H species and the activated Ni–formate complex. To substantiate the claimed bimetallic hydride transfer, the authors should incorporate in situ techniques such as IR, paramagnetic NMR, or provide a detailed justification for the absence of this data.

We have attempted several in situ control experiments, however, in our mechanistic studies we face two main limitations: (1) the paramagnetism of [Cp*Co(4DHBP)]I and Ni(acac)₂ complicating NMR studies and (2) the decomposition of [Cp*Co(4DHBP)]I at concentrations higher than (approximately) 2 mM under catalytic conditions.

Despite the paramagnetic character we explored NMR techniques for in situ experiments. When [Cp*Co(4DHBP)]I is pressurized in a NMR tube with 40 bar of H₂ in presence of D₂O and NaHCO₃, the only peaks detected are small peaks corresponding to the fraction of disassociated 4DHBP ligands.  No peak corresponding to the complex or the eventual Co-H are detected, despite the solution maintaining its distinctive purple color and forming no particulate. We propose the formation of a cobalt-hydride given the fact that [Cp*Co(4DHBP)]I is active towards formate production even without the presence of Ni(acac)₂ under basic conditions. We also base our hypothesis on the DFT calculations carried by Himeda et al. and Yang et al. (ref. 42 and 40 in the main text) on similar cobalt complexes.

The second technique explored for mechanistic studies is high-pressure FT-IR. While [Cp*Co(4DHBP)]I concentrations lower than 1 mM do not allow for signal detection, concentrations higher than 2 mM afford a dark precipitate.

 

 

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

accept