Pyridine-2,6-Dicarboxylic Acid Esters (pydicR2) as O,N,O-Pincer Ligands in CuII Complexes
, Maryam Heydari Dokoohaki 2, Amin Reza Zolghadr 2,* and Axel Klein 1,2,*Round 1
Reviewer 1 Report
The authors report preparation and a detailed characterization of several new neutral and anionic Cu(II) complexes derived from 6-pyridinedicarboxylic acid and its methyl, phenyl and 2-iodophenyl esters, ‘pydicR2’ (R = Me, Ph, C6H4I). Some of the products are characterized by single crystal X-ray diffraction. The coordination geometry of the Cu(II) center, charge separation and H-bonding interactions between counterions in ionic alkylammonium cuprates(II) are discussed based on results of DFT calculations, EPR and UV-VIS spectroscopic characterization.
The results are clearly presented and may be of interest to inorganic chemists working in the field of copper chemistry. The manuscript may be published after a minor revision described below.
1) In Scheme 2 triethylamine and 4-dimethylaminopyridine are designated as ‘catalysts’ whereas they are reagents. Please correct.
Author Response
Reviewer 1:
R1.1: In Scheme 2 triethylamine and 4-dimethylaminopyridine are designated as ‘catalysts’ whereas they are reagents. Please correct.
Response: Scheme 2 was redrawn
Reviewer 2 Report
The authors present a study of a series of tridentate ONO ligands and the metallated Cu(II) analogs in monomeric and dimeric forms. The authors discuss the synthesis and spectroscopy from an experimental perspective and add to this with some electronic structure discussion based on DFT with the addition of NBO and some MD work. As far as I can tell the authors made one and only one new Cu complex. In addition their one-pot synthesis might be new, but not surprising.
Overall the work is scientifically sound and presented well. I think that Inorganics is a good target for this manuscript as the paper is not really significant or novel and I don't expect the broad interest you would need for higher tier journals.
Overall I see no issues with the science, but I do however have many issues, both large and small, with the presentation of the information, as well as some suggestions about how the authors could simply improve the manuscript.
First, I wish that the authors expanded on and collect their discussion of the theoretical methods. The DFT is discussed in a very limited way in section 3 with the other experimental information, but the MD details are buried in the body of the paper. In addition the discussion of the DFT methods was too limited. The authors did calculations in both turbomole and gaussian, but it is not clear why they switched platforms. In addition the calculations in turbomole and gaussian used different basis sets. I assume the rationale for this is what was available on each platform, but the authors never distinguish the results in the paper as having been from one method or the other. This is not acceptable. You can't compare these numbers and put them in the same tables without noting which ones are which, and really you shouldn't be doing it at all. The authors also performed NBO calculations, but didn't mention how they performed these calculations at all. I assume it was in Gaussian. All of this must be documented and cited in the experimental sections.
As for the content, there are a few aspects to the discussion that make me question whether the authors know anything at all about electronic structure or if they are just parroting words they have heard before. The most glaring case is the repeated misuse of the Jahn-Teller effect and it's results. JT distortions occur ONLY when you have degenerate orbitals with unequal occupations. This would happen for instance in an Oh Cu(II) were the eg orbital has 3 electrons. The authors, however, discuss JT distortions in their C2v symmetric pseudo-octahedral compounds! THIS CANNOT HAPPEN!!! In the C2v point group these are not degenerate orbitals! In this pseudo octahedral structure the a1 dx2-y2 (which is really z2-x2) is likely the SOMO (the authors could tell us since they have done the DFT) but it isn't degenerate with the a1 dz2 (or more accurately the dy2). There is no energetic benefit in JT sense to distorting along the z axis because that wouldn't lower the symmetry and break the degeneracy. Now, that doesn't mean that the axial Cl bonds aren't longer than the other one, but if it isn't it isn't JT because that's not what JT is. Nonsense like this distracts from the entire paper. I'll give them the benefit of the doubt and allow them to just remove all reference to JT here and call that enough.
This fundamental lack of knowledge about electronic structure and the underlying group theory is apparent when the authors mention visible bands that correspond to the t2g -> eg transition. WRONG! There are no orbitals of either t2g or eg symmetry in any of these molecules. Using Oh labels for a pseudo-octahedral molecule is wrong. Again, the authors have the DFT generated Kohn-Sham orbitals, they could tell us about them and point to the orbitals they think are involved in these transitions. They could even compare the energy gaps between the d-orbitals to the absorbance bands. Better yet the could model the spectroscopy with TDDFT, but that's is quite a bit more sophisticated then what they have already done. Again, I would settle for the authors to just state things correctly. These transitions are likely from the primarily non-bonding d-orbitals of "t2g parentage" (xy, xz and yz) or d z2 into the singly occupied d x2-y2 (Assuming that that is in fact the SOMO). It is not acceptable to use Oh symmetry for C2v molecules.
The authors correctly included solvent-optimized calculations in the manuscript and it seems that their not surprising conclusion was that this is important. What was not clear, however, was if the different solvent systems provided any interesting insight. The authors state on p 9 that these were important, and on p 10 that the differences were "not that high". I guess you could argue that the fact that the differences were not significant is in itself significant. If that is the case then state it. Here, though, I question whether the authors have gone far enough. For MeCN and DMF did the authors ever try to use explicit solvent molecules. Do they know if these molecules can bind to the open coordination site on the neutral 5-coordinate Cu? I don't think the MD calculations can help out with this as you would really need a reactive force field. There are things that can be done if the authors wanted to pursue them.
My last large issue is that the authors start the conclusion section off by stating New CuII complexes ... were synthesised...
Were they? Really? The authors state on p 3 that one of the I complexes was new, but as far as I can tell everything else has been reported before. That would mean there is only 1 new complex.
There are a couple of small places where the wording is strange. On p 7, I don't understand the use of "Besides" to open up the paragraph. Besides what? That word isn't used correctly. The authors also use the phrase "more aquatic" in the abstract. This is nonsense. If they mean aqueous instead of aquatic that gets them half way there, but that's not what they're doing either. They're making things more polar, but they're not making things more aqueous and they sure aren't doing anything aquatic, that's a macroscopic term relating to large biological systems.
Overall I think the paper is good enough for Inorganics but the conclusions seem to only be that they made 1 new compound and that using a solvent model matters.
Author Response
Reviewer 2:
R2.1: First, I wish that the authors expanded on and collect their discussion of the theoretical methods. The DFT is discussed in a very limited way in section 3 with the other experimental information, but the MD details are buried in the body of the paper. In addition the discussion of the DFT methods was too limited.
Response: We have outlined the MD details a bit more in the Experimental Section 3.3.
R2.2: The authors did calculations in both turbomole and gaussian, but it is not clear why they switched platforms. In addition the calculations in turbomole and gaussian used different basis sets. I assume the rationale for this is what was available on each platform, but the authors never distinguish the results in the paper as having been from one method or the other. This is not acceptable. You can't compare these numbers and put them in the same tables without noting which ones are which, and really you shouldn't be doing it at all. The authors also performed NBO calculations, but didn't mention how they performed these calculations at all. I assume it was in Gaussian. All of this must be documented and cited in the experimental sections.
Response: The choice of choosing TMOLE vs. Gaussian is of pure technical nature (availability). In the Experimental Section 3.3 we have outlined that the question of a mononuclear [Cu(pydic(IPh)2)Cl2] vs. a binuclear [{Cu(pydic(IPh)2)Cl}2(µ-Cl)2] structure was tackled using TMOLE and def-SV(P)/B3LYP basis sets. For the calculations on the two compounds (HNEt3)[Cu(pydicR2)Cl3] (R = Me or Ph) we used Gaussian and B3LYP/6–31+G(d,p) basis sets. Since we do not compare the pydic(IPh)2 complex with the latter, the use of different platforms and basis sets should not be a problem. We have revised the text and tables to make this clearer. In Tables S2 to S7 this is correctly described.
R2.3: As for the content, there are a few aspects to the discussion that make me question whether the authors know anything at all about electronic structure or if they are just parroting words they have heard before. The most glaring case is the repeated misuse of the Jahn-Teller effect and it's results. JT distortions occur ONLY when you have degenerate orbitals with unequal occupations. This would happen for instance in an Oh Cu(II) were the eg orbital has 3 electrons. The authors, however, discuss JT distortions in their C2v symmetric pseudo-octahedral compounds! THIS CANNOT HAPPEN!!! In the C2v point group these are not degenerate orbitals! In this pseudo octahedral structure the a1 dx2-y2 (which is really z2-x2) is likely the SOMO (the authors could tell us since they have done the DFT) but it isn't degenerate with the a1 dz2 (or more accurately the dy2). There is no energetic benefit in JT sense to distorting along the z axis because that wouldn't lower the symmetry and break the degeneracy. Now, that doesn't mean that the axial Cl bonds aren't longer than the other one, but if it isn't it isn't JT because that's not what JT is. Nonsense like this distracts from the entire paper. I'll give them the benefit of the doubt and allow them to just remove all reference to JT here and call that enough.
Response: We have removed the term “Jahn-Teller” and replaced it by descriptive terms like “axially elongated”. The reviewer is correct with its criticism that only relieving the eg degeneracy justifies the assignment of a Jahn-Teller distortion and it cannot be simply concluded from bond elongation or shortening, most of all not in non-octahedral systems like ours.
On the other hand, in the review by Malcolm A. Halrow (ref. 47) for a C2v symmetry the d levels are noted: a1 (dx2-y2) > a1 (dz2) (non-degenerate) > a2 (dxy) > b1 (dyz) = b2 (dxz) (degenerate) and we stick to this notation although putting the z-axis as the main molecular axis should be preferred from a spectroscopic point of view.
R2.4: This fundamental lack of knowledge about electronic structure and the underlying group theory is apparent when the authors mention visible bands that correspond to the t2g -> eg transition. WRONG! There are no orbitals of either t2g or eg symmetry in any of these molecules. Using Oh labels for a pseudo-octahedral molecule is wrong. Again, the authors have the DFT generated Kohn-Sham orbitals, they could tell us about them and point to the orbitals they think are involved in these transitions. They could even compare the energy gaps between the d-orbitals to the absorbance bands. Better yet the could model the spectroscopy with TDDFT, but that's is quite a bit more sophisticated then what they have already done. Again, I would settle for the authors to just state things correctly. These transitions are likely from the primarily non-bonding d-orbitals of "t2g parentage" (xy, xz and yz) or d z2 into the singly occupied d x2-y2 (Assuming that that is in fact the SOMO). It is not acceptable to use Oh symmetry for C2v molecules.
Response: The reviewer is correct and we have corrected this: … broad bands assignable to d-d transitions (a2,b1,b2→a1; in C2v symmetry) were observed … . TD-DFT calculations would indeed largely exceed the scope of this report (synthesis, structures, EPR, basic UV-vis absorption spectroscopy and electrochemistry).
R2.5: The authors correctly included solvent-optimized calculations in the manuscript and it seems that their not surprising conclusion was that this is important. What was not clear, however, was if the different solvent systems provided any interesting insight. The authors state on p 9 that these were important, and on p 10 that the differences were "not that high". I guess you could argue that the fact that the differences were not significant is in itself significant. If that is the case then state it. Here, though, I question whether the authors have gone far enough. For MeCN and DMF did the authors ever try to use explicit solvent molecules. Do they know if these molecules can bind to the open coordination site on the neutral 5-coordinate Cu? I don't think the MD calculations can help out with this as you would really need a reactive force field. There are things that can be done if the authors wanted to pursue them.
Response: On page 10, we have replaced “not very high” by “are not very pronounced”. We disagree that “the differences were not significant” (see e.g. Table 2) and we also do not want to write something “less than expected”.
We have indeed discussed the option of doing MD calculations using explicit solvent molecules and check for coordination or ligand exchange. We refrained from doing so, since that would be a large effort and even more importantly, we have no experimental evidence for any association or dissociation and could thus not compare the results of such calculations with experimental findings.
R2.6: My last large issue is that the authors start the conclusion section off by stating New CuII complexes ... were synthesised... Were they? Really? The authors state on p 3 that one of the I complexes was new, but as far as I can tell everything else has been reported before. That would mean there is only 1 new complex.
Response: In this manuscript we report three new compounds (HNEt3)[Cu(pydicMe2)Cl3], (HNEt3)[Cu(pydicPh2)Cl3], and[{Cu(pydic(IPh)2)Cl}2(µ-Cl)2]. We have rephrased some parts to make this clearer and have redrawn Scheme 2.
R2.7: There are a couple of small places where the wording is strange. On p 7, I don't understand the use of "Besides" to open up the paragraph. Besides what? That word isn't used correctly. The authors also use the phrase "more aquatic" in the abstract. This is nonsense. If they mean aqueous instead of aquatic that gets them half way there, but that's not what they're doing either. They're making things more polar, but they're not making things more aqueous and they sure aren't doing anything aquatic, that's a macroscopic term relating to large biological systems.
Response: A) We simply deleted “Besides”, the sentence starts no with “To better understand … . B) “more aquatic” has been replaced by “in the presence of water”.
Finally, we want to express that we are very grateful to reviewer 2 for the critical evaluation of our manuscript helping us to avoid sluggish terminology.
