Chiral-Icosahedral (I) Symmetry in Ubiquitous Metallic Cluster

Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Chiral-Icosahedral (I) Symmetry in Ubiquitous Metallic Cluster Compounds (145A,60X): Structure and Bonding Principles Published as part of the Accounts of Chemical Research special issue “Toward Atomic Precision in Nanoscience”. Robert L. Whetten,*,† Hans-Christian Weissker,‡,§,¶ J. Jesús Pelayo,∥ Sean M. Mullins,† Xochitl López-Lozano,† and Ignacio L. Garzón*,⊥ †

Department of Physics & Astronomy, University of Texas, San Antonio, Texas 78249, United States Aix Marseille University and European Theoretical Spectroscopy Facility, CNRS, CINaM UMR 7325, 13288 Marseille, France § European Theoretical Spectroscopy Facility ∥ Escuela Superior de Apan, Universidad Autónoma del Estado de Hidalgo, Chimalpa Tlalayote, Apan, 43920 Hidalgo, México ⊥ Instituto de Física, Universidad Nacional Autónoma de México, Apartado Postal 20-364, 01000 CDMX, México

Acc. Chem. Res. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/03/19. For personal use only.

‡

CONSPECTUS: There exists a special kind of perfectionin symmetry, simplicity, and stabilityattainable for structures generated from precisely 60 ligands (all of a single type) that protect 145 metal-atom sites. The symmetry in question is icosahedral (Ih), generally, and chiral icosahedral (I) in particular. A 60-fold equivalence of the ligands is the smallest number to allow this kind of perfection. Known cluster compounds that approximate this structural ideal include palladium-carbonyls, Ih-Pd145(CO)60; gold-thiolates, IAu144(SR)60; and gold-alkynyls, I-Au144(C2R)60. Many other variants are suspected. The Pd145 compound established the basic achiral structure-type. However, the Au144-thiolate archetype is prominent, historically in its abundance and ease of preparation and handling, in its proliferation in many laboratories and application areas, and ultimately in the intrinsic chirality of its geometrical structure and organization of its bonding network or connectivity. As discovered by mass spectrometry (the “30-k anomaly”) in 1995, it appeared as a broad single peak, as solitary and symmetrical as Mount Fuji, centered near 30 kDa (∼150 Au atoms), provoking these thoughts: Surely this phenomenon requires a unique explanation. It appears to be the Buckminsterfullerene (carbon-60) of gold-cluster chemistry. Herein we provide an elementary account of the unexpected discovery, in which the Pd145-structure played a critical role, that led to the identification and prediction, in 2008, of a fascinating new molecular structure-type, evidently the first one of chiral icosahedral symmetry. Rigorous confirmation of this prediction occurred in early spring 2018, when two single-crystal X-ray crystallography reports were submitted, each one distinguishing both enantiomeric structures and noting profound chirality for the surface (ligand) layer. The emphasis here is on the structure and bonding principles and how these have been elucidated. Our aim has been to present this story in simplest terms, consistent with the radical simplicity of the structure itself. Because it combines intrinsic profound chirality, at several levels, with the highest possible symmetry-type (icosahedral), the structure may attract broader interest also from educators, especially if studied in tandem with the analysis of hollow (shell) metallic systems that exhibit the same chirality and symmetry. Because the shortest (stiffest) bonds follow the chiral 3-way weave pattern of the traditional South-Asian reed football, this cultural artifact may be used to introduce chiral-icosahedral symmetry in a pleasant and memorable way. One may also appreciate easily the bonding and excitations in I-symmetry metallic nanostructures via the golden fullerenes, that is, the proposed hollow Au60,72 spheres. Beyond any aesthetic or pedagogical value, we aim that our Account may provide a firm foundation upon which others may address open questions and the opportunities they present. This Account can scarcely hint at the prospects for further fundamental understanding of these compounds, as well as a continued...

Received: September 23, 2018

© XXXX American Chemical Society

A

DOI: 10.1021/acs.accounts.8b00481 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. (a) Escher’s rendition of icosahedral symmetry.1 (b) Escher’s “Drawing hands” represents nonsuperimposable mirror images or enantiomeric forms.1 (c) The traditional South-Asian football combines these elements; it is centered between two enantiomeric models thereof. Panels a and b are reprinted from ref 1. M. C. Escher’s “Gravitation” and “Drawing Hands”. Copyright 2018 The M.C. Escher Company, The Netherlands. All rights reserved. www.mcescher.com. Panel c is adapted with permission from ref 9. Copyright 2013 PCCP Owner Societies.

Figure 2. (a) Ih-Pd145 experimental structure.5 (b) Au144(SCH2Ph)60 experimental structure.3 (c) I-Au144(SCH3)60 theoretically predicted structure31,35 (d) Au144(phenyl-alkynyl)60 experimental structure.4 Metal atoms on C5 axes (red), on C2 axes (inner orange; outer gold), and not on rotational axes (blue). Distribution of atoms from the center of mass (e) and from nearest neighbors (f) for the experimental3 and predicted31,35 Au144(thiolate)60 compounds. The inset in panel f depicts the trans configuration of the staple-motif. The bond lengths and radial distances were convoluted with a Gaussian, σ = 0.01 eV. The theoretical data was obtained using the fully symmetric structure31 relaxed using LDA.

Figure 3. (a−d) Representation of the achiral−chiral transition, or symmetry-breaking, by use of the transition between the two relevant polyhedra (lower left, b−d). Orbital energy levels (e−f), in eV, of chiral gold shells Au60 and Au72, and frontier orbitals of Au60 (g-h). Reproduced with permission from ref 8. Copyright 2018 Nature Communications (under a Creative Commons Attribution 4.0 International License).

widening sphere of applications (chemical, electronic, imaging). The compounds remain crucial to a wider field presently under intense development.

B

DOI: 10.1021/acs.accounts.8b00481 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research 1.2. Surprising PdCO-AuSR Connection

1. HISTORY OF THE I-(145A,60X) COMPOUNDS One may ask: Are there new principles of molecular structure and bonding that await discovery via research on atomically precise clusters? Here, we gather evidence for chiral-icosahedral (I) structures in the ubiquitous noble-metal clusters comprising 145 metal-atom sites and 60 anionic ligands. In the classification of molecular structures, icosahedral symmetry ranks highest. It is defined by 21 axes of rotation {6 C5, 10 C3, 15 C2} (Figure 1a).1 But chirality involves broken (lowered) symmetry, that is, nonsuperimposable mirror images (Figure 1b).1 I-symmetric structures integrate these opposing elements. Earlier discussions of potential I-symmetry molecules describe these as unprecedented, as a “difficult problem in stereochemistry”.2 Then how remarkable it is that such structures arise naturally! New research on the abundant 145site metallic clusters, archetypically the gold−thiolate compounds I-Au144(SR)60, has confirmed earlier speculations of a simple I-symmetry motif, based upon single-crystal X-ray diffractometry.3,4 These structures resemble the achiral Ih-Pd145 structure (Figure 2a).5 All exhibit concentric shells of {1@12@(30 + 12)@60@30} metal-atom A-sites as well as 60 bridging ligand X-sites, (145A,60X). But in key respects they differ. The Au144 clusters display enhanced nobility and metallicity,6,7 a central vacancy, an intrinsically chiral bonding network, and immense variability in ligands and metals (alloys). Geometrically, the coordinating sites (60A,60X) resemble respective achiral and chiral Archimedean (60-vertex) polyhedral forms, Figure 3.8 The shortest (strongest) bonds follow the chiral 3-way weave pattern of the traditional South-Asian football (Figure 1c).9 Here, we recount the discovery of the Au144-thiolate compounds,10 their connection to the Ih-Pd145 structure-type, the analytical barriers to their characterization and realistic modeling,11 and the recent establishment of the structure.3,4 Theoretical works predicted the structures to a remarkable precision, including their chiral-icosahedral (I)-symmetry.12,13 Simplified pictures of their electronic structure and bonding allow chemists to comprehend the principles applicable to other ligand-protected noble-metal clusters and SAMs generally, but are best exemplified in cases of highest symmetry and simplest organization.14−16 Among fundamental properties described are the electrical (electrochemical) phenomenon of “molecular capacitance”, the unusually discrete optical absorption spectra, and related indications of metallic bonding character.17,18 Finally, we devote special attention to the analysis of the chirality of (145A,60X) compounds: its origin, its measure (magnitude), and its diverse implications.

By early 1999, single crystals were obtained for a palladium (Pd) compound, Pd145(CO)60(PEt3)30.5 The Pd145 structure revealed (Figure 2a) an essentially perfect icosahedral symmetry: concentric shells of 1@12@{30 + 12}@60 equivalent metalatom sites form a Pd115 core; 30 external Pd atoms coordinated to the R3P: ligands; and 60 carbonyls (CO) bridge the shell of 60 Pd atoms. The relevance of this wondrous Pd145(CO)60(PEt3)30 cluster to the mysterious gold−thiolates was unclear: Pd (group 10, 4d105s0) and Au (group 11, 5d106s1) clusters are distinctive, as are their respective coordination chemistries with carbonyl (:CO) and thiolate (-SR, anionic) ligands.22 Yet the Pd145 atomic coordinates yielded an astoundingly good match to the measured X-ray scattering function of the ∼30 kDa gold compound.6 Previously proposed models had all failed this simple test. Thus, the hypothesis: the mystery compound’s formula must be “Ih-Au145(SR)60”.6 1.3. Analytical Challenge

But “Ih-Au145(SR)60” compounds remained refractory to precise determination of composition and structure. During the period 1997−2006, many laboratories prepared gold−thiolate ∼30 kDa substances, frequently adopting their own nomenclature, based on imprecise estimates of ligand and metal-atom numbers. (Similar confusion affected analysis of smaller clusters.)23 The uncertainty hardly affected work directed toward applications, which depended more on facility of synthesis and purification; tolerance to air, moisture, light, and heat; and ease of modification. The analytical problem lays in deficiencies of extant methods for solving anything as complex as “Au145(SR)60”, notwithstanding the prior structure determination of Pd145 (CO)60 . Consequently, major advances came first on smaller homologues. By 2005, high-resolution ESI-MS work on clusters, separated by gel electrophoresis,24 established precise compositions for several smaller gold−thiolate compounds, for example, Au25SG18, where GSH is the tripeptide glutathione. 1.4. Crowding Relieved by Chirality

One objection to the “Au145(SR)60” hypothesis is that there is simply too little space for 60 n-alkyl-thiolates to fit around the globular Au115 core of icosahedral (Ih) symmetry. But a relaxation from full icosahedral (Ih) to chiral-icosahedral (I) symmetry relieves the steric “crowding” to allow four more ligands, 56 vs 60. To see this, compare the two Archimedean polyhedra, Figure 3,8 wherein 60 vertices represent 60 α-CH2 groups. Assuming a fixed distance separating groups, the surface area required by 60 vertices (ligands) is 55.28 units for the chiral structure vs 59.30 for the achiral, an increase of +7.3%. This geometrical insight suggested that chirality is intrinsic to the structure. In 2007, the Stanford biophysical group14 had solved the crystal structure of a larger gold−thiolate compound (102,44), wherein RS- = p-mercaptobenzoic acid (pMBA). By determining the location of all heavy atoms (Au, S, C, O), one finally obtained a detailed picture of how the surface bonding is compatible with a large (79-site) core that resembles bulk metallic gold.24

1.1. Basic Observations

The earliest recorded encounter occurred in 1995,19 by way of a single-phase variant of the Brust−Schiffrin 2-phase method,20 generating larger gold-rich thiolated clusters. Naturally, the mass spectrum showed a single peak, centered near 30-kDa (∼150 Au atoms). The assignment of a molecular formula requires stronger evidence of composition, which arrived slowly until 2008, when the precise ligand-count (60) was finally established. Early assessments (elemental analysis, X-ray scattering, microscopy) indicated a Au/S ratio of ∼2.4:1 and ordered metallic core bonding throughout a core of diameter ∼16 Å, which agrees with mass 30 kDa assuming a globular shape and bulk density.18,21 Spectroscopy (IR, NMR) identified the intact thiolate (RS) ligands.

1.5. Composition (144,60) Established

To test a complete structure model, one final piece of the puzzle was required: In 2008, Chaki et al.11 reported mass spectrometry (ESI-MS) of a series of the “30 kDa” gold−alkanethiolate C

DOI: 10.1021/acs.accounts.8b00481 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research samples. Use of strongest peaks yields the formula 144-Au, 59thiolates (144,59). However, weaker peaks at slightly higher masses yield (144,60), that is, 60-ligand protection of an icosahedral core structure.11 Subsequent work confirmed the (144,60) composition.25,26

2. STRUCTURE AND BONDING IN THE I-(145A,60X) COMPOUNDS 2.1. Complete Model Refined

Combining the (144,60) composition, the Pd145 frame, and the “staple-motif” coordination,24 one arrives at a unique model of the type I-Au145(SR)60 amenable to refinement by state-of-theart DFT electronic structure calculations.12 Starting from the available Pd145 coordinates,5 one inserts 30 pairs of bridging thiolates, each pair joined collinearly via an external Au(I) site, and each S atom linked radially to a surface Au atom [cf. Figure 2]. Each staple motif spans diagonally a rectangular edge of the Ih-Au115 grand core. The construction confirms the following extraordinary fact: There exists only one way (two, counting enantiomers) to accomplish this construction. The structure is intrinsically chiral, not only for the steric reasons (above) but as demanded by coordination of Au(I)−thiolate staple motifs. In refining the structure,12 it was found to be more stable (less strained) minus the central atom, that is, one of the 145 sites remains vacant. The model was then optimized at the highest level then feasible and reported along with many predictions.12 Over the next decade, this “Standard Model” accounted for numerous experimental results on A144−145X60 systems (A = Au, Ag, Cu, or Pd; X = thiolate variants),27 some mentioned below.28−30

Figure 4. Electronic structure of the chiral-icosahedral I-Au144X60 compounds is represented by the orbital energy levels (horizontal axes, in eV units) and the number of orbitals (degeneracy, or density of states, DOS) at each energy level (vertical axis). See text for details. Top panel, isoelectronically substituted Au144Cl60 obtained with the SIESTA code.9 Middle panel is adapted with permission from ref 35. Copyright 2015 American Chemical Society. The angular-momentum projected DOS in the bottom panel was obtained by projecting Kohn−Sham wave functions of the fully symmetrized structure from a ground state calculation with the real-space code Octopus onto spherical harmonics. Bottom panel is adapted with permission from ref 38. Copyright 2017 The Royal Society of Chemistry.

2.2. I-Symmetrization

In 2012, the compound Au144(SCH2CH2Ph)60 was used as a “target” to demonstrate rapid-scanning single-molecule structure determination by selected-area electron diffraction.31 This necessitated construction of an all-trans (symmetrized) version of the Standard Model,12 a task initiated by Tlahuice-Flores9 and reported by Bahena et al.31 The symmetrized structure was further simplified, via iso-electronic (halide) substitution, for elucidation of the structure-bonding principles.9 Two key lessons are (i) exact I-symmetry holds whenever the chargestate fills an electronic shell [cf. Figure 4] and (ii) a set of 60 exceptionally short Au−Au radial bonds complete a simple sixstranded weave pattern, Figure 1c. Cryogenic EXAFS results soon confirmed the “bond stiffness” (rigidity) of these shortest intergold bonds and the entire weave pattern.28 Extension of the symmetrized (all-trans) structures to more realistic ligands and Au−Ag−Cu alloys preserves I-symmetry and frontier-orbital character, assuming the preferred charge states. This subsequent work, led by Weissker et al.,30,35 in parallel to improvements by Fortunelli et al.32 and Malola et al.,33 advanced confidence in the “Standard Model”.34

These are essential steps toward certain practical goals, namely, (i) to achieve a fundamental understanding commensurate to the broader significance of the compounds-as-utilized, (ii) to provide guidance to future high-level experimental work that confirms and complements such theoretical comprehension, and (iii) to reassure those who prefer simply to use such compounds that the fundamentals are soundly understood. The following subsections describe highlights of this progress, and section 3 focuses on the special issue of chirality. 2.4. Geometrical and Electronic Structure Comparison

In Figure 2(b,c,e,f), we compare the experimental structure, with R = CH2Ph,3,35 with an earlier prediction using R = CH3 and the LDA functional, which for the similarly sized Au146(pMBA)57 produced excellent bond lengths.36 The predicted connectivity and orientation (all-trans) of the staples and therefore the fully symmetrized structure are confirmed,30 with profound consequences for the electronic and optical properties.30 Spectacular agreement is seen for the distances from the center of mass (central vacancy), showing shells of {12; 30 + 12; 60; 30} symmetry-equivalent gold atoms, and for the interatomic distances.

2.3. Totally Determined Structures and Implications

However, chemists generally expect to see at least one case of rigorous total-structure determination of a I-Au144X60 compound, by single-crystal X-ray crystallography. In early Spring (2018), two such reports were submitted to journals.3,4 Each distinguished both enantiomeric structures and noted profound chirality for the surface (ligand) layer. The abundant detail in such crystallographic data sets provides rich material for comparison to predictions of theoretical models and thus to stimulate further refinements.

2.5. Electronic and Optical Properties

The electronic and optical properties of the Au144 compound stand out amid the larger gold−thiolate clusters. Figure 4 compares the strongly peaked (symmetry-related) DOS of the D

DOI: 10.1021/acs.accounts.8b00481 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research fully symmetrized “Bahena” model30,31,35 and the isoelectronically substituted model of ref 9 with that of the less symmetrical “Lopez-Acevedo” structure,35 that is, the model that reported first the correct connectivity but lower symmetry due to cis orientation of several staple motifs.12 The frontier orbitals of the globular cluster follow the superatom coordination-complex model (SAC).34 The globular metallic core supports frontier states that have clean angularmomentum (L) character, (bottom panel in Figure 4), much as in atomic electronic structure. The chiral-icosahedral field does, however, produce characteristic splittings that greatly reduce the HOMO−LUMO gap.12,37,38 Low-temperature absorption spectra of Au144 compounds show rich, individually peaked spectral features in the visible and near-infrared regions, much more so than any similarly sized system.29,30 The spectral information relates to the electronic structure, making Au144 convenient to explore the transition from molecular spectra in smaller, molecule-like clusters to larger nanoparticles with smooth electronic bands and smooth spectra. The rich spectral information makes it likewise a precious benchmark system for the necessary improvement of calculations, concerning both the proper approximations of TDDFT and the high sensitivity to geometrical details of the models.35 The plasmonic resonance only starts to emerge in this size range and is fully evident at ∼300 gold atoms.33 Figure 5 compares spectra calculated for two structures shown in Figure 2b,c, one obtained theoretically (R = CH3, fully

3. CHIRALITY IN Au144X60 AND RELATED COMPOUNDS: MEASURES AND ORIGIN Chirality in thiolate-protected gold clusters (TPGCs), Aun(SR)m, was discovered in 1998,39,40 based upon intense optical activity in the visible spectrum of gold−glutathione clusters with size