Reverse Self-Assembly: (111)-Oriented Gold Crystallization at Alkylthiol Monolayer Templates

It has long been known that thiol-terminated molecules self-assemble as commensurate monolayers on Au(111) surfaces. By spreading floating octadecanethiol monolayers on aqueous solutions of chloroauric acid (${\mathrm{HAuCl}}_4$) and using x rays to reduce the gold ions as well as to probe the structure, we have observed the nucleation of (111)-oriented Au nanoparticles at thiol surfaces. This process may be similar to the formation of biogenic gold by bacteria. The thiol monolayer acts as a “soft template,” changing its structure as Au crystals form so that there is a $\sqrt{3}\times \sqrt{3}$ commensurate relationship.

Figure 1

A representation of our “reverse self-assembly” experiment at the air-water interface. (a) The upper half of each frame shows a floating monolayer formed by spreading octadecanethiol on an aqueous solution of ${\mathrm{HAuCl}}_4$. (b) The synchrotron beam produces elemental gold by photoreduction and is also used to characterize the resulting gold crystals through diffraction.

Figure 2

(a),(b) In situ GID intensity contours of gold crystals. The peak positions confirm that the gold particles are (111)-oriented. (c) Time resolved one-dimensional diffraction scans derived from $(\overline{2}20)$ peak in (b). Data are integrated over $Q_z=0.02–0.04{\AA }^{-1}$.

Figure 3

Electron micrographs of gold crystals collected from the air-water interface. (a) TEM micrograph after $\sim 20\min$. (b),(c) Scanning electron microscope micrographs after 2 and 4 h, respectively. It is possible to see the hexagonal plate at the base of the flowerlike particle in (c).

Figure 4

(a),(b) In situ GID intensity contours of octadecanethiol peaks at two different times, before and after gold crystals formed. (c) Peak due to the molecular bilayer, which does not move with time. (d)–(g) One-dimensional diffraction scans derived from the intensity contours. Data are integrated over $Q_z=0.02–0.04{\AA }^{-1}$ for the horizontal scan (d) and $\Delta Q_{xy}=\pm 0.01{\AA }^{-1}$ around the given $Q_{xy}$ value for the vertical scans (e)–(g). The data for 10 and 100 min have different scales for better visibility, in all plots except (d).