Evolution from the plasmon to exciton state in ligand-protected atomically precise gold nanoparticles

The evolution from the metallic (or plasmonic) to molecular state in metal nanoparticles constitutes a central question in nanoscience research because of its importance in revealing the origin of metallic bonding and offering fundamental insights into the birth of surface plasmon resonance. Previous research has not been able to probe the transition due to the unavailability of atomically precise nanoparticles in the 1–3 nm size regime. Herein, we investigate the transition by performing ultrafast spectroscopic studies on atomically precise thiolate-protected Au25, Au38, Au144, Au333, Au∼520 and Au∼940 nanoparticles. Our results clearly map out three distinct states: metallic (size larger than Au333, that is, larger than 2.3 nm), transition regime (between Au333 and Au144, that is, 2.3–1.7 nm) and non-metallic or excitonic state (smaller than Au144, that is, smaller than 1.7 nm). The transition also impacts the catalytic properties as demonstrated in both carbon monoxide oxidation and electrocatalytic oxidation of alcohol.

The electron dynamics show a qualitative change from molecule-like for small clusters (with bleaching at discrete transition energies, excited-state absorption, and single decay times) to metal-like for large clusters (with a broadened and shifted plasmon resonance and an electron cooling time that depends on the energy of the exciting pulse). Au_333 exhibit a combination of both features, and are thus describes as being intermediate between the two regimes. This size also coincides with the point where catalytic activity no longer increases with cluster size, but rather begins to decrease.
These results address the question of how plasmonic behavior emerges in gold nanocrystals as their size increases. This, in turn, is an excellent test bed for the greater question of how bulk behavior emerges from the interactions among atoms, a grand challenge in condensed-matter physics and materials science. Although the limiting behaviors have previously been observed, the current manuscript provides a systematic study, enabled by the spectacular ability to synthesize highly monodisperse colloidal gold clusters over a broad range of sizes. The observation of an intermediate size regime, in particular, provides new insight into the microscopic-macroscopic crossover, and should inspire new theoretical studies. I thus consider the manuscript to be novel and important, and deserving of publication in Nature Communications. I have only a few points that the authors should take into consideration.
1. The results are specific to the ligand-stabilized colloidal gold clusters that the authors synthesize. This is important to emphasize, because the quantitative crossover is undoubtedly different in these clusters than in ligand-free gas-phase clusters, or indeed in other types of ligand-stabilized metal clusters. Especially for the smaller clusters, changes in cluster size are accompanied by changes in geometry that may have an impact on their electronic properties and the molecular-metallic crossover. I would suggest that the authors revise the abstract to clarify the types of ligand-stabilized clusters that are studied, and perhaps add the words "ligand-stabilized" and/or "colloidal" to the title.
2. The authors correctly point out the importance of choosing the correct probe wavelength when extracting time constants from transient-absorption measurements, particularly in the case of the Au_144 clusters. This ambiguity could be avoided by performing a global analysis on the data. If there is indeed a single time constant, it should be able to extract this time constant from a global analysis based on singular-value-decomposition (relatively common in analysis of transient-absorption data), and a more precise estimate of the time constants may be obtained in this way.
3. As a minor point, the manuscript appears to mostly refer to Au_333 clusters as intermediate between molecular and metallic, but at some points seems to refer to them as metallic. It would be best to be consistent throughout the text.
4. I also note some minor grammatical errors throughout the text. The revised version should be carefully proofread for correct usage.

Reviewer #2 (Remarks to the Author):
Transition from molecular to metallic energetics was probed by ultrafast spectroscopy analysis in ligand stabilized gold nanoclusters. This is a very important question to address in this fast growing field and has broad implications in chemistry, physics and materials science. The authors used the same stabilizing ligand to synthesize a series of gold clusters with 'atomic' precision, at least for the smaller ones. Previous related studies are limited by the size distribution and incoherence in ligand/S-bonding, both addressed in this report. The qualitative agreements and discrepancies with those in literature (i.e. by Goodson, Knapenberger as cited among others) will likely stimulate discussions and future work. I would recommend publication after some minor revisions.
Discussions on why such large clusters like Au144 with multiple shells is nonmetallic could provide physical insights for generalization and broaden the scope of the work. Scan rate in CV should be provided. The peak potentials should be compared to other catalysts under the same alkaline conditions.

Author Reply:
We have added that "The CV scan rate is 20 mV s -1 " in the method part (page 11). The main purpose of this manuscript is to compare the catalytic behavior of nanoparticles in the transition regime. It would not be fair to compare with other catalysts since their composition, size, and surface ligand are quite different.
I will be curious to see the catalytic efficiency normalized by outer shell Au atoms. By weight is acceptable though.

Reviewer #3 (Remarks to the Author):
This manuscript reports studies of the ultrafast electron dynamics and catalytic activity of monodisperse, ligand-stabilized, colloidal gold clusters with sizes from 25 to approximately 950 atoms. The electron dynamics show a qualitative change from molecule-like for small clusters (with bleaching at discrete transition energies, excited-state absorption, and single decay times) to metal-like for large clusters (with a broadened and shifted plasmon resonance and an electron cooling time that depends on the energy of the exciting pulse). Au_333 exhibit a combination of both features, and are thus describes as being intermediate between the two regimes. This size also coincides with the point where catalytic activity no longer increases with cluster size, but rather begins to decrease.
These results address the question of how plasmonic behavior emerges in gold nanocrystals as their size increases. This, in turn, is an excellent test bed for the greater question of how bulk behavior emerges from the interactions among atoms, a grand challenge in condensed-matter physics and materials science. Although the limiting behaviors have previously been observed, the current manuscript provides a systematic study, enabled by the spectacular ability to synthesize highly This is important to emphasize, because the quantitative crossover is undoubtedly different in these clusters than in ligand-free gas-phase clusters, or indeed in other types of ligand-stabilized metal clusters. Especially for the smaller clusters, changes in cluster size are accompanied by changes in geometry that may have an impact on their electronic properties and the molecular-metallic crossover. I would suggest that the authors revise the abstract to clarify the types of ligand-stabilized clusters that are studied, and perhaps add the words "ligand-stabilized" and/or "colloidal" to the title.
Author Reply: We have followed the reviewer's suggestion and indicated the presence of ligands.
Revision: In the abstract, we have added "thiolate-protected". In the title, we have corrected as "ligand-protected, atomically precise gold nanoparticles".