Single-gold etching at the hypercarbon atom of C-centred hexagold(I) clusters protected by chiral N-heterocyclic carbenes

Chemical etching of nano-sized metal clusters at the atomic level has a high potential for creating metal number-specific structures and functions that are difficult to achieve with bottom-up synthesis methods. In particular, precisely etching metal atoms one by one from nonmetallic element-centred metal clusters and elucidating the relationship between their well-defined structures, and chemical and physical properties will facilitate future materials design for metal clusters. Here we report the single-gold etching at a hypercarbon centre in gold(I) clusters. Specifically, C-centred hexagold(I) clusters protected by chiral N-heterocyclic carbenes are etched with bisphosphine to yield C-centred pentagold(I) (CAuI5) clusters. The CAuI5 clusters exhibit an unusually large bathochromic shift in luminescence, which is reproduced theoretically. The etching mechanism is experimentally and theoretically suggested to be a tandem dissociation-association-elimination pathway. Furthermore, the vacant site of the central carbon of the CAuI5 cluster can accommodate AuCl, allowing for post-functionalisation of the C-centred gold(I) clusters.

The manuscript effectively presents a comprehensive overview of rationally conducted experiments including SCXRD and theoretical calculations to show the changes at the atomic level.Nevertheless, I have reservations regarding the paper's suitability for publication in Nature Communications.To enhance its compatibility with this prestigious journal, it would greatly benefit from a more thorough reflection on the novelty of the work making it align with that required for Nature Communications, as it is not very clear in the introduction section.Authors have published work related to NHC-protected CAu16 clusters and in this work, they investigate whether etching occurs when bisphosphine is added to NHC-protected CAu16 clusters.While similar etching SS-1NHC and 1TPP with 1,2-bis(diphenlnylphosphino)benzene also yielded the corresponding CAu15 clusters (ref 26, and 46 in the manuscript).Authors should include a chart showing all the clusters and their abbreviated labels to help readers navigate easily, and the literature compounds should be referenced.The significance of the low flack parameter for SS, RR-2NHC should be mentioned briefly.
Reviewer #2 (Remarks to the Author): The manuscript, "Single-gold etching at the hypercarbon atom of C-centred hexagold(I) clusters protected by chiral N-heterocyclic carbenes", is a comprehensive report on the photophysical properties of CAu(I)5 cluster and mechanism of chemical etching method to reduce one Au(I) atom from CAu(I)6 cluster.DFT calculations were performed to interpret the properties and to reinforce the proposed mechanism.This paper has a strong impact on how we can synthesize metal cluster by manipulating one-by-one chemical reaction.The unusual cluster-size dependence of the photophytical property is also intriguing.I think this manuscript is worth publishing in Nature Comm after minor comments are considered.No further review is necessary.
In abstract, 4th line, this sentence is hard to read.Please revise.
Is the same procedure applicable to form CAu(I)4 cluster?The CAu(I)4 cluster is charge neutral, and four Au(I) can fill the valency of the central C atom.
In the hexagonal and pentagonal clusters, the carbon atom should be in a very unusual electronic state.Please give a brief explanation.
The description about theoretical calculation is not enough and hard to reproduce.Please consider reproducibility.
In the TDDFT calculation, B3LYP functional was used.This functional is not a long-range corrected one, and the description of charge-transfer type excitation should be carefully verified.In this case, TDDFT calculation reproduced the observed spectrum, and the performance of the functional was verified.But, I still feel that the authors give some comments on the B3LYP functional for application to the MLCT type excited states of the small Au clusters.
The unusual red-shift of the CAu(I)5 cluster from the CAu(I)6 cluster was explained on the basis of the HOMO-LUMO gap.The HOMO contains the C component, and the energy level shows significant highenergy shift.I would like the authors to give an explanation on the reason of the shift.The work is extremely well done and presented .This contribution should be well received by the Nature Comm readership.I have a few general questions: 1.Why are chiral NHCs used here -would achiral NHCs also work to demonstrate the generality of this surgical method?2. if AuCl can be added as a cap to the Au octahedron could other metals be added as well?now this would open a new frontier in the field as luminescence properties could possibly be tuned in multimetallic systems.I am curious -could CuCl be added or other transition metal fragments?
Possibilities and questions abound.So an excellent contribution.
A few specific comments: 1. line 33: have made .... 2.line 43-are still a matter of debate 3. line 44 for surgically modifying metal clusters 4. line 51 -cannot be has made progress better: has attracted attention and significant advances have been made in this area.

in fig 2 show structure of depe
Reviewer #4 (Remarks to the Author): I co-reviewed this manuscript with one of the reviewers who provided the listed reports.This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co-review manuscripts.
Responses to Reviewer #1 ABSTRACT: Etching is an excellent top-down synthesis method for controlling the structures and chemical and physical properties of nanomaterials.Chemical etching at the atomic level is a particularly challenging task for the precise synthesis of nanosized metal clusters.The synthesis of metal clusters containing smaller metal nuclei and potential surface vacancies will further elucidate the details of structure-function relationships and facilitate future materials design.Here we report the successful single-gold etching at a hypercarbon centre in gold(I) clusters.Specifically, C-centred hexagold(I) clusters protected by chiral N-heterocyclic carbenes were etched with bisphosphine to yield C-centred pentagold(I) (CAuI5) clusters.The CAuI5 clusters exhibit an unusually large bathochromic shift in luminescence, which is reproduced theoretically.The etching mechanism was experimentally and theoretically suggested to be a tandem dissociation-association-elimination pathway.Furthermore, the vacant site of the central carbon of the CAuI5 cluster can accommodate AuCl, allowing for post-functionalisation of the C-centred gold(I) clusters.Pei et al report the "Single-gold etching at the hypercarbon atom of C-centred hexagold(I) clusters protected by chiral N-heterocyclic carbenes" The manuscript effectively presents a comprehensive overview of rationally conducted experiments including SCXRD and theoretical calculations to show the changes at the atomic level.Nevertheless, I have reservations regarding the paper's suitability for publication in Nature Communications.To enhance its compatibility with this prestigious journal, it would greatly benefit from a more thorough reflection on the novelty of the work making it align with that required for Nature Communications, as it is not very clear in the introduction section.

Response:
We thank you for your important comments that helped us improve the manuscript.In order to better emphasize the significance and novelty of the chemical etching method developed in this study, we have added a few statements to the "Introduction" and "Discussion" sections as shown below.

Revision in the manuscript Page 3, "Introduction" section, paragraph 1 Before revision
Etching is an excellent top-down method to subtractively modify the structures and chemical and physical properties of a wide range of nanomaterials such as nanocrystals and colloidal nanoparticles 1-5 , and nanoclusters 6-11 for a variety of applications (Fig. 1a).

(omission)
The ligand-exchange mechanisms of aggregative bimolecular nucleophilic substitution (SN2)like or unimolecular nucleophilic substitution (SN1)-like type in nanogold regions containing Au I and Au 0 atoms are still a matter of debate [20][21][22] .Despite the promise of chemical etching as a universal technique for precisely "surgical" metal clusters at the atomic level, little attention has been paid to ligand-protected Au I clusters.

After revision
Etching is an excellent top-down method to downsize the structures at the atomic level and modify the chemical and physical properties of a wide range of nanomaterials such as nanocrystals and colloidal nanoparticles 1-5 , and nanoclusters 6-11 for a variety of applications (Fig. 1a).(omission) It also remains controversial whether the ligand-exchange mechanisms in nanogold regions containing Au I and Au 0 atoms is SN2-like bimolecular nucleophilic substitution or SN1-like type unimolecular nucleophilic substitution [20][21][22] .Despite the promise of chemical etching as a general technique to downsize metal clusters at the atomic level, little attention has been paid to ligand-protected Au I clusters.

After revision
In particular, the hypercoordinated carbon (hypercarbon) 35 -centred hexagold(I) (CAu I 6) cluster [C(Au I -L)6]X2 (L = ligand; X = counterion) 26 that bridges nano-sized metal clusters and organic molecules 36,37 has attracted attention and significant advances have been made in this area.Page 4, "Introduction" section, paragraph 3 Before revision Here we have found that C-centred hexagold(I) clusters protected by chiral NHCs can be etched with a bisphosphine ligand to yield chiral NHC-protected CAu I 5 clusters, and showed that the gold(I) clusters can be precisely operated at the atomic level by controlling the number of gold atoms centred at the hypercarbon (Fig. 1b).Furthermore, this etching method also works for the synthesis of TPP-protected CAu I 5 analogues.Although there is a general rule that smaller gold clusters show more blue-shifted emission than larger clusters, the ligand-protected CAu I 5 clusters show unusually red-shifted signals in both absorption and emission spectra compared to the CAu I 6 counterparts, which was rationalised by theoretical calculations.Further experimental and theoretical studies suggest that the etching mechanism involves a tandem pathway of dissociation-association-elimination.Moreover, the reactivity of the CAu I 5 clusters at the hypercarbon centre was further confirmed by the addition of AuCl.

After revision
Here, we discovered that a chiral NHC-protected C-centred hexagold(I) cluster can be etched with a bisphosphine ligand to generate a chiral NHC-protected CAu I 5 cluster.This was achieved by controlling the number of gold atoms centred at the hypercarbon at the atomic level (Fig. 1b).Furthermore, this etching method is also useful for the synthesis of TPPprotected CAu I 5 analogues.In general, smaller gold clusters show more blue-shifted emission than larger clusters, but the ligand-protected CAu I 5 clusters show unusually red-shifted signals in both absorption and emission spectra compared to the CAu I 6 counterparts, which was rationalised by theoretical calculations.Further experimental and theoretical studies suggest that a tandem dissociation-association-elimination pathway is involved in the etching mechanism.The NHC-protected CAu I 6 clusters are generally chemically stable, and the active site had to be placed on the hypercarbon to confer reactivity.In fact, CAu I 5 synthesised by this method was highly reactive with AuCl, producing a CAu I 6 cluster with a different ligand.Thus, the chemical etching method is expected to be an excellent way not only to reduce the size of metal ion clusters and significantly change their electronic structure, but also to asymmetrise the metal ion cluster structure and provide active sites.
Page 18, "Discussion" section, last sentence Before revision In particular, we can expect to gain new knowledge about reaction mechanisms such as addition reactions, elimination reactions, and catalytic reactions of organic compounds and their correlations.

After revision
Therefore, the chemical etching method is an excellent way to reduce the size of metal ion clusters and will be developed to control of the electronic structure, asymmetrisation of the metal ion cluster structure, catalytic reactions, and metal ion exchange.
Authors have published work related to NHC-protected CAu16 clusters and in this work, they investigate whether etching occurs when bisphosphine is added to NHCprotected CAu16 clusters.While similar etching SS-1NHC and 1TPP with 1,2bis(diphenlnylphosphino)benzene also yielded the corresponding CAu15 clusters(ref 26, and 46 in the manuscript).Authors should include a chart showing all the clusters and their abbreviated labels to help readers navigate easily, and the literature compounds should be referenced.
Response: To make it easier for the readers to distinguish between all the clusters described in this paper, the compounds in the references are shown in Fig. 2 and the schematic diagram in Supplementary Fig. 1.

(omission)
Revision in the Supplementary Information electronic state.Please give a brief explanation.

Response:
The unique electronic structure of the central carbon atom is of interest.The orbital interactions have been discussed in previous studies 50,51 .In this study, the bond orders of C-Au I and Au I ⋅⋅⋅Au I were analysed in detail as described in the main text (p.9-10).For simplicity, the following sentence was added at the end of the 4th paragraph in the "Results" section.

Revision in the manuscript
9, "Results" section, the end of paragraph 4 Before revision (omission) Therefore, missing one Au atom in the CAu I 5 cores may result in stronger Ccentre-Au I bonds, which may be important for stabilising the CAu I 5 clusters.After revision (omission) Therefore, missing one Au atom in the CAu I 5 cores may result in stronger Ccentre-Au I bonds, which may be important for stabilising the CAu I 5 clusters.Moreover, the orbital interactions of the CAu I 6 cluster were previously discussed in detail 50,51 : the SS-1 NHC and SS-2 NHC clusters have [CAu6] 2+ and [CAu5] + cores, respectively, and the C-Au I bond orders exhibit an unusual C-Au I bond hypervalence.
The description about theoretical calculation is not enough and hard to reproduce.Please consider reproducibility.
Response: Details of the theoretical calculations are described in the "Theoretical calculation details for etching mechanism" section in the Supplementary Information on page S7 so that the results can be reproduced.

Revision in the Supplementary Information
Page S7, "Theoretical calculation details for etching mechanism" section Before revision The proposed intermediates including Int1 NHC , Int2 NHC , Int1 TPP and Int2 TPP were simulated and optimised using the Gaussian 16 suite of programs 7 .The calculated energy profiles were illustrated in Supplementary Fig. 32.

After revision
The proposed intermediates including Int1 NHC , Int2 NHC , Int1 TPP and Int2 TPP were simulated and optimised using the Gaussian 16 suite of programs 7 .Optimisation was performed using the B3LYP functional combined with basis sets of 6-31G* (for C, N, P, H) and LANL2DZ (for Au).The solvent effects were evaluated by single-point calculations in the optimised structure using the polarisable continuum model (PCM).All chemical species involved were optimised in the singlet state.The calculated energy profiles were illustrated in Supplementary Fig. 33.
In the TDDFT calculation, B3LYP functional was used.This functional is not a longrange corrected one, and the description of charge-transfer type excitation should be carefully verified.In this case, TDDFT calculation reproduced the observed spectrum, and the performance of the functional was verified.But, I still feel that the authors give some comments on the B3LYP functional for application to the MLCT type excited states of the small Au clusters.

Response:
We appreciate the reviewer for addressing this point.As reviewer noted, the MLCTs in the present small Au clusters do not show long-distance charge transfer and they essentially correspond to the charge reorganization and therefore, global hybrid functionals like B3LYP suitably describe these electronic transitions.We added this explanation in Supplementary Information (page S4).

Revision in the Supplementary Information
Page S4, "DFT and TD-DFT calculations" section Before revision We applied the B3LYP functional 3 for geometry optimizations and TD-DFT calculations.The r elativistic effective core potential LANL2DZ 4 was used for the Au atoms, and the basis set for the other atoms was 6-31G* 5 .For simulating absorption spectra, (omission)

After revision
We applied the B3LYP functional 3 for geometry optimisations and TD-DFT calculations.The relativistic effective core potential LANL2DZ 4 was used for the Au atoms, and the basis set for the other atoms was 6-31G* 5 .Since MLCTs in the Au clusters etched in this study do not show long-distance charge transfer and essentially correspond to charge reorganisation, global hybrid functionals such as B3LYP adequately describe these electronic transitions.For simulating absorption spectra, (omission) The unusual red-shift of the CAu(I)5 cluster from the CAu(I)6 cluster was explained on the basis of the HOMO-LUMO gap.The HOMO contains the C component, and the energy level shows significant high-energy shift.I would like the authors to give an explanation on the reason of the shift.

Response:
The significant energy shift of the orbital energy levels is attributed to the total charge of the cluster.Indeed, the carbon component in the HOMO of the CAu I 5 cluster (32.3%) is larger than that of the CAu I 6 cluster (24.6%) (ref.43, SI).The CAu I 5 and CAu I 6 clusters have total charges of +1 and +2, respectively, which leads to the large orbital energy difference or shifts between these clusters; the IPs (ionization potentials) of C and Au are 11.2 and 9.2 eV, respectively.Considering the orbital interactions, the ratio of carbon in the HOMO cannot explain the significant energy shift.Response: The Int2 L (L = NHC, TPP) shown in Fig. 4f has only one cis-depe ligand, while the product of compound 3 [Au(cis-depe)2(BF4)] has two cis-depe ligands.This is because a slight excess of cis-depe (2.5 equiv) was added to the dichloromethane solution of the CAu I 6 cluster (1 equiv).After removal of [cis-depeAu] + in Int2 L (L = NHC, TPP), the remaining cis-depe in solution would bind with [cis-depeAu] + to form the stable species of compound 3. To clarify this point, Fig. 4f was modified as follows.Response: We would like to thank you for your very positive comments.Our response to each comment is as follows.

Revision in the manuscript
To comment 1. line 33: have made ... Page 3, "Introduction" section, paragraph 1 Revised according to the reviewer's suggestion.Before revision For example, chemical etching methods that involve ligand engineering made great advances, After revision For example, chemical etching methods that involve ligand engineering have made great advances, To comment 2. line 43-are still a matter of debate Page 3, "Introduction" section, paragraph 1 In order to better emphasize the significance and novelty of the chemical etching method developed in this study, we have added a few statements to the "Introduction" section as shown below.

Before revision
Etching is an excellent top-down method to subtractively modify the structures and chemical and physical properties of a wide range of nanomaterials such as nanocrystals and colloidal nanoparticles 1-5 , and nanoclusters 6-11 for a variety of applications (Fig. 1a).

(omission)
The ligand-exchange mechanisms of aggregative bimolecular nucleophilic substitution (SN2)like or unimolecular nucleophilic substitution (SN1)-like type in nanogold regions containing Au I and Au 0 atoms are still a matter of debate [20][21][22] .Despite the promise of chemical etching as a universal technique for precisely "surgical" metal clusters at the atomic level, little attention has been paid to ligand-protected Au I clusters.

After revision
Etching is an excellent top-down method to downsize the structures at the atomic level and modify the chemical and physical properties of a wide range of nanomaterials such as nanocrystals and colloidal nanoparticles 1-5 , and nanoclusters 6-11 for a variety of applications (Fig. 1a).(omission) It also remains controversial whether the ligand-exchange mechanisms in nanogold regions containing Au I and Au 0 atoms is SN2-like bimolecular nucleophilic substitution or SN1-like type unimolecular nucleophilic substitution [20][21][22] .Despite the promise of chemical etching as a general technique to downsize metal clusters at the atomic level, little attention has been paid to ligand-protected Au I clusters.
To comment 3. line 44 for surgically modifying metal clusters Please see our response to your comment 2.
To comment 4. line 51 -cannot be has made progress better: has attracted attention and significant advances have been made in this area.Revised according to the reviewer's suggestion.

Figure 4f ,
Figure 4f, Int2L should have one depe ligand, but the two depe ligands are eliminating in compound 3.

Figure 4f ,
Figure 4f, Int2L should have one depe ligand, but the two depe ligands are eliminating in compound 3.