Nano-Golden Order

See allHide authors and affiliations

Science  19 Oct 2007:
Vol. 318, Issue 5849, pp. 407-408
DOI: 10.1126/science.1150176

Iridescent gold nanostructures, exploited since antiquity in the medical, decorative, and theological arts, have at last begun to succumb to total structure determination, the sine qua non of the molecular sciences. On page 430 of this issue, Jadzinsky et al. report a breakthrough in this field: the total structure determination of a large gold-thiolate cluster (1). The work will help to nail down the detailed structure of self-assembled monolayers (SAMs), which are chemically modified gold electrodes with applications from sensing to nanolithography (2).

Atomic structures of ever larger molecules are determined by x-ray crystallography, but structural studies of nanometer-scale metal clusters (3) or of surfaces are still challenging. In the case of clusters, creating identical clusters can be difficult, although important exceptions have been reported (4). Surface diffraction studies also encounter numerous resolution-limiting problems, such as large contributions from subsurface metal layers and multiple scattering effects.

The molecules used to create SAMs on gold surfaces (see the figure, top panel) consist of an organic group (R) linked via a sulfur atom (S) to gold substrate atoms (Au). Extended SAMs on crystalline substrates are typically highly ordered, allowing the orientation of the R group to be determined, but the structure of the buried head group is often only inferred. However, structural models built on the basis of such inferences (see the figure, top panel) have been hotly contested.

Because chemical bonding is largely a local affair, one approach to attacking this nanostructure problem is to use gold nanocrystals as the metal substrate (see the figure, bottom left) (4, 5), on the assumption that the gold-thiolate bond is constant at all scales and that a suitable sample will crystallize for total structure determination. Massive thiolate monolayer-protected gold clusters of various dimensions and with different R groups have been produced, usually from decomposition of the corresponding nonmetallic Au(I)-thiolate precursors (see the figure, bottom right). The chemical and electrical properties of these clusters usually closely resemble those of the extended SAMs. They remain exceptionally stable during size separation and analysis, stimulating claims of molecular levels of homogeneity, but until now no total structure determination has been achieved (6).

How do thiolates bind to gold?

Several binding modes for self-assembled monolayers have been suggested (top). Gold-sulfur binding modes in the 102-gold-atom cluster reported by Jadzinsky et al. (1) (bottom left) may reveal the structure of self-assembled monolayers. Gold(I)-thiolate compounds have also been structurally characterized (bottom right). In all panels, gold atoms are depicted as yellow circles, sulfur as orange circles; the hexagons denote a phenyl (-C6H5) aromatic hydrocarbon.

Enter Jadzinsky et al., who report the successful isolation, crystallization, and structure determination of a cluster with 102 gold atoms and 44 RS groups (see the figure, bottom left). They accomplished this feat by using water-soluble, rigid aromatic benzene thiolates (7) with a carboxylate group (not shown in the figure) on the opposite side of the aromatic ring from the sulfur atom. Crystallization trials with the reaction products yielded large single crystals suitable for x-ray diffraction measurements. Given the great effort previously expended to deduce structural aspects of such clusters, it is a relief to find that the structural features of the cluster are mostly familiar (8), although they are combined in ways that could not have been guessed.

Seventy-nine of the 102 gold atoms form a truncated-decahedral cluster. This “grand core” has gold-gold bond lengths and coordination consistent with bulk metals, supported gold nanocrystals, and a prior model for gold-thiolate clusters ∼1.5 nm in diameter (9). In an unexpected twist, the 15-atom caps at each pole are rotated with respect to the ideal compact decahedral shells, yielding denser packing and placing all 40 surface atoms of the grand core on the surface of a common sphere.

The remaining gold atoms form an outer shell that interacts strongly with the thiolate sulfur atoms. Jadzinsky et al. obtained a clear picture of this interaction (the “surface-chemical bond”). The results are in striking contrast to the standard model (see the figure, top), in which independent thiyl groups are attached directly to close-packed gold substrates. Rather, the binding motif exemplifies the “divide-and-protect” model (10), in which oligomeric aurous Au(I)-thiolate complexes interact weakly with the surface atoms of the grand core (see the figure, bottom left). Surface species of this type have analogs in known anionic complexes (see the figure, bottom right) (11, 12).

The weak interactions of the thiolates with the surface atoms of the grand core are least well understood. A strongly polar, if not ionic, interaction is suggested by the distinctive anchoring of thiolate sulfur atoms in formally anionic complexes (10, 11) to single substrate gold atoms in a correspondingly positive core. The close approach of Au(I) atoms within the Au(SR)2 adsorbates to the surface atoms of the grand core is consistent with relativistically enhanced dispersion interactions (13). To gain additional insight, thermodynamic measurements and first-principles theory are crucially needed.

It is intriguing that the cluster's composition—with 102 gold atoms and 44 thiolates—suggests an additional electronic origin for its stability. In metal-cluster physics, electronic rules of cluster stability are well established. According to these rules, filled concentric angular-momentum shells of electrons confer electronic structural stability to a cluster. One such filled shell requires 58 electrons to occupy 29 delocalized orbitals. This number is achieved exactly if each of the 102 gold atoms donates one electron to delocalized orbitals, and if each of the 44 thiyl radicals (RS˙) formally takes one electron into a localized orbital. Electronic rules of cluster stability have been used widely in gas-phase elemental clusters, but only rarely in molecular cluster chemistry (14); the current results suggest that they should perhaps be applied more widely.

It is fortuitous that a simple procedure (reductive degradation of oligomeric precursors) yields a self-assembled, discrete compound (15) that contains structural organization consistent with known precedents and classification (11). The known properties of nanoscale clusters can now be rationalized in terms of atomic ordering. Do structures of this type hold generally for flatter clusters and extended SAMs, for smaller clusters with more highly curved surfaces, for non-aromatic R groups, and for diverse gold-plated and gold-alloy nanostructures? If so, then there is hardly a published interpretation (for example, of electron transfer, capacitance, or density) that will not be in need of revision or reinterpretation.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
View Abstract

Stay Connected to Science

Navigate This Article