Computational Investigation of Nickel-Mediated B–H Activation and Regioselective Cage B–C(sp2) Coupling of o-Carborane
Round 1
Reviewer 1 Report
Mu, Hao and co-workers presents a comprehensive computational study on the mechanism of an interesting reaction, the multicomponent nickel-catalyzed activation of o-carborane and its coupling with an alkene and a silyl-alkyne. The mechanism is complex and could help in the further development of similar reactions. Also, the regioselectivity of this process is also explored, although a predictive model, that could help in the improvement of the diasteromeric ratio, is not provided and would be a valuable addition to the study. In addition, I have few concerns about the methodology employed (they used several different functionals but the quality of the basis set is not enough to claim for accurate results). For all of this, I recommend publishing the paper after major revisions (see below):
Major revisions:
The authors put enough effort to describe the differences among different functionals, also including implicit solvation. However, due to the multicomponent character of the reaction, the basis set can influence the overall results and the one used in the study (DZVP) is far from the state of the art and can produce a significant BSSE. For this reason, I believe that the calculation single points with a bigger basis set (triple-zeta level) would improve the quality of the results (see recent literature as: ACS Catal. 2015, 5, 5545−5555; Science 2019, 363, 391–396). This can be done in a reasonable time and would improve the manuscript.
In several complexes, the nickel coordination sphere is unsaturated. According to scheme 1, the nickel precursor used in the experimental reaction contains two phosphines. The authors should check the stability of those unsaturated intermediates upon coordination of the second phosphine (M1, INT1, INT4, INT5, TS5, COM2, TS6, INT6...). This could change considerably the relative barriers and the resting state of the catalytic cycle.
Minor revision:
The authors selected as the starting point the intermediate COM1 instead of the initial species M1. I do not understand this: M1 is more reasonable and gives information about the alkyne coordination to the system.
The barriers for the carbonate free reaction are above 45 kcal/mol, too high for a reaction at 110ºC with a 79% yield. This probably is because of the BSSE and the absence of dispersion in the main calculations. I would expect a more reasonable numbers after the correction of the basis set and the consideration of the second phosphine as assisting ligand.
Aryl bromides and aryl iodide can be activated by nickel catalysis via single electron transfer, followed by radical recombination. This possibility should be tested in the reaction from INT6 (see for example: Chem. Eur. J. 2017, 23,16728–16733 and Angew .Chem. Int. Ed. 2019, 58, 3898–3902).
Author Response
The authors would like to thank both reviewers for careful review of our manuscript and providing us with helpful comments and suggestions to improve the quality of our manuscript. Those comments are very valuable and significant for polishing this paper. We really appreciate your precious time and great efforts on reviewing the manuscript. We have carefully revised our manuscript taking your comments and suggestions into account, and the following responses were prepared to address all of your comments.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The topic of the work is in the field of interest, however, there are some points mainly based on the computational methods what should be overviewed and re-concluded.
On the other hand several things should be also changed to represent the work in a more logical way.
At first the studied o-carborane should be drawn in a scheme because it is not visible (but obvious) which atoms are boron and carbon as well. I suggest a scheme of the zirconium complex to be drawn. R1, R2 and R3 also should be mentioned in the text of the introduction previously, not only in the figure legend.
There are extra spaces in the line 177 and 239.
In the line 268 it written “many an experiment…” which is confusing.
In the line 315 arrows should be replaced.
Figure 2 and 4 are hardly visible and therefore they are not relevant and could be transferred to the SI.
Figure 3 and 6 can be drawn together, however, in the whole work it is a bit hard to observe the atom B11. I prefer to use a (similar) symbol such as for carbon atoms.
Table 2 should be discussed before Table 3, and INTs should be also included in the tables.
Main comments:
I do not understand the name of COM1 and COM2, and I disagree that the energetic study is based on the reference structure COM1. This is the first intermediate (INT1) of the pathway derived from the M1 + R3, however, M1 is also an intermediate coming from the zirconium pre-catalyst. Therefore I would like to see the energetics of the reaction R1 + [Ni] → M1 + rest, and to set the reference at the energy sum of the M1 + R2 + R3. I declare that choosing a reference is most of the time a problematic point of any mechanistic study.
I do not prefer to call triple bond in the COM1, however, the lengthening is not significant compared to R3, but it is a coordinated substrate which has a delocalized electron system.
There is a transition state of the isomerization of INT1 and INT2 what should be found, and it is incorrect to state that TS2 has a barrier of 16.9 kcal mol–1 because it is 15.5 kcal mol–1 starting from INT2.
As I see INT6 is the form of M3.
It is not obvious how the relative energy of M2 was compared to COM2. I guess it just coming from COM2 → M2 + [Ni], however, it is not written in the Fig. S1.
And it is also not clear how the energies have been re-calculated by other functionals. Were these geometry optimizations or just single point calculations?
Cesium has been replaced by potassium what can cause a huge effect on energies. I know that Cs is not trivial to describe, however, I suggest to use CRENBL ECP for cesium (J Mass Spectrom. 2015 Jan;50(1):240-6.) or Stuttgart-Dresden relativistic ECP (Journal of Molecular Structure 1148 (2017) 206e212). I believe that calculations with cesium can improve the novelty of this work.
In line 247 it is written that “Path a, in which Cs2CO3 directly participates in the B-H activation” which is not right due to the fact that there is no cesium in the calculations.
I do not see the relevance of the calculated reaction rate constant.
I do not think that CAM-B3LYP is a good choice for studying the energetics of a mechanism, it is more useful for simulating excitation states as this statement is also supported by the results.
The huge difference between B3LYP and B3LYP-D3 makes the reader confused because according to results finally it is not clear why B3LYP results are represented in the figures. It is obvious that dispersion and long-range corrections are necessary to be taken into account.
I guess to re-calculate the mechanism at B3LYP-D3/TZVP and LC-wPBE/TZVP levels because between DZVP and TZVP the difference can be significant, and the description of the nickel atom is also needed to be more accurate. It would be also interesting to see the difference between B3LYP-D3 and M06 to check the effect of the dispersion correction in different ways.
In the conclusion LC-wPBE was chosen as the best functional according to the calculated diastereoselective ratio (dr, P1b:P1a), however, I guess it can be changed significantly by using TZVP basis sets.
I feel that the first part of the paper is very well written but later there are trivial sentences and some repeats what decreases the level of the description of the results.
I believe that with a major revision and some important changes the impact of this work, and it can be published at Catalysts.
Author Response
The authors would like to thank both reviewers for careful review of our manuscript and providing us with helpful comments and suggestions to improve the quality of our manuscript. Those comments are very valuable and significant for polishing this paper. We really appreciate your precious time and great efforts on reviewing the manuscript. We have carefully revised our manuscript taking your comments and suggestions into account, and the following responses were prepared to address all of your comments.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors improved the manuscript according to my suggestions, so I recommend to accept the manuscript.
The only point that could be still improved (because it was misunderstood by the authors) is the radical pathway that they commented in page R5 of the response. They excluded the pathway because of the high energy of the Br radical and the Ni(I)-Aryl intermediate. However, this is not the way that the bond would break homolitically. This bond would break in a Ni(I)-Br fragment and the more stable aryl radical.
I would appreciate the correction of this part but I am fine with the current state of the paper and I do not need to review again the manuscript.
Author Response
The authors would like to thank both reviewers again, for your hard work and precious comments on our manuscript, which are very helpful and significant for improving the quality of this paper. We have carefully revised our manuscript by taking all the comments and suggestions into account, and the following responses were prepared to address all of your points.
Author Response File: Author Response.pdf
Reviewer 2 Report
The authors have tried to minimalize the efforts to improve this nice work, however, I really believe that publishing B3LYP results is not sufficient, especially with DZVP basis sets.
I have a doubt about the comparison to the work of Hostas and Rezac. They have written about the accuracy of DZVP in association with dispersion corrected functionals for geometry optimizations and the related energies, however, TZVP optimizations can cause larger differences, mainly during TS optimizations (and for activation barriers) due to the fact that the reaction coordinates are shifted by using larger basis sets. Not surprisingly TZVP single point calculations have not given large effect on energies.
The significant difference of the results calculated by different methods have not been discussed in details, and it is obvious that LC-ωPBE or B3LYP-D3 result more accurate data. On the other hand authors should prove their comment "probably arisen from overweight of the weak interaction between KHCO3, KBr and other fragments" somehow by any citation etc. I like the calculations (and to see the obvious differences) in the case of cesium but I do not understand why the results of potassium are written in the paper.
In this present form I do not like this work mainly because of the represented B3LYP/DZVP results and the conclusion, respectively.
Comments for author File: Comments.pdf
Author Response
The authors would like to thank both reviewers again, for your hard work and precious comments on our manuscript, which are very helpful and significant for improving the quality of this paper. We have carefully revised our manuscript by taking all the comments and suggestions into account, and the following responses were prepared to address all of your points.
Author Response File: Author Response.pdf
