Gold nanoparticle ring arrays from core–satellite nanostructures made to order by hydrogen bond interactions

Polyethylene glycol-grafted gold nanoparticles are attached to silica nanoparticle cores via hydrogen bonding in a controlled fashion, forming well-defined core–satellite structures in colloidal solution. For separating these complex structures effectively from the parental nanoparticles, a straightforward and easy protocol using glass beads has been developed. The attached gold nanoparticles show unique surface mobility on the silica core surface, which allows for nanoparticle rearrangement into a 2D ring pattern surrounding the silica nanoparticle template when the core–satellite structures are cast to a planar surface. When etching away the silica core under conditions in which the polymer shell fixes the satellites to the substrate, highly ordered ring-shaped patterns of gold nanoparticles are formed. By variation of the size of the parental particles – 13 to 28 nm for gold nanoparticles and 39 to 62 nm for silica nanoparticles – a great library of different ring-structures regarding size and particle number is accessible with relative ease. The proposed protocol is low-cost and can easily be scaled up. It moreover demonstrates the power of hydrogen bonds in polymers as a dynamic anchoring tool for creating nanoclusters with rearrangement ability. We believe that this concept constitutes a powerful strategy for the development of new and innovative nanostructures.

2 of 39 nm SiNPs (120 μL, 0.1 mg/mL in THF) in a 1.5 mL glass vial under bath sonication. The mixture was allowed to be further sonicated for additional 30 minutes and then placed on an orbital shaker (150 rpm) overnight to ensure a saturation of AuPEG on the SiNPs surface. The detailed feed ratios for different SiAu-3D nanoclusters prepared in this work are summarized in Tab. S1.
Purification of SiAu-3D by glass beads. Based on our experience, the purification performance is highly dependent on the amount ratio between the glass beads and excess AuPEG rather than the concentration of the colloid. In this way, we can estimate the amount of glass beads for each type of AuPEG with different AuNPs sizes. We further found that the Ø 0.5 mm glass beads have the most balanced performance and handling property thus suitable for all types of AuPEG used in this work. In general, 12, 4, and 5 mg glass beads are sufficient to remove 1 μg of 13, 21, 28 nm-sized AuNPs (the weight of PEG shell is ignored here).
In a typical run, glass beads were added in two steps to the unpurified nanocluster colloid. Extra THF was added to keep all glass beads below the liquid level. After each addition step of the glass beads, the sample was placed on an orbital shaker (175 rpm) for 20 minutes. The purified colloid was separated from the glass beads to a clean glass vial by a micropipette. TEM characterization was employed to control the purification result. If any free AuPEG remains, an additional purification step can be executed by using a decreased amount of glass beads to achieve the ideal purification result.
Notably, for the sample with shorter PEG chains (800 g/mol, 13 nm AuNPs), only a 25% dosing of glass beads is required to completely remove all AuPEG.
Formation of SiAu-2D on substrate. To bring SiAu nanocluster from colloid onto a substrate, we simply drop-cast 10 μL of SiAu-3D colloid onto a carbon film-covered TEM grid. The solvent was allowed to evaporate under ambient conditions. For silica removal application, lacey carbon film was used instead of the common carbon film for better liquid penetration which yields more consistent etching results. For AFM characterization, 2 μL of SiAu-3D colloid was drop-cast on a piece of gold-covered silicon wafer. The substrates were covered with a glass vial and left undisturbed under ambient conditions overnight to ensure a complete evaporation of the solvent.
Selective removal of SiNP template. TEM grid (lacey carbon film) carrying SiAu-2D at the surface was incubated in a 10 wt% aqueous NaOH solution overnight. After incubation, the TEM grid was gently dipped in water and ethanol successively to remove any remaining NaOH. The TEM grid was then left undisturbed for solvent evaporation before TEM characterization.
Characterizations. TEM imaging was performed on a Philips CM 12 electron microscope. The TEM was operated at an acceleration voltage of 120 kV and an emission current of 3-4 µA. The DLS was performed on a Malvern Zetasizer Nano S system operating at 633 nm at 25 °C. Samples were measured after a 120 s equilibration period within the analyzer without previous filtration. UV-Vis absorption spectroscopy was performed with a Jasco V770 scan photo spectrometer in solution against pure solvent (baseline subtraction method). A halogen lamp was used as the light source. AFM measurement was carried out on a Bruker Multimode 8 equipped with a Scanasyst-Air-HR cantilever. Fig. S1 (A-C) TEM images of 6000 g/mol PEG capped AuNPs with diameters of (A) 13 ± 1 nm, (B) 21 ± 2 nm, and (C) 28 ± 3 nm, respectively. (D) 13 nm AuNPs capped with 800 g/mol PEG.  Tab. S1 Detailed experimental parameters for self-assembly process of SiAu nanoclusters with different SiNPs and AuPEGs. For the calculation, the density of AuNPs is 19.3 g/cm 3 , the density of SiNP is 2.2 g/cm 3 . The diameter of each NP is determined from TEM statistically ( Fig. S1-S2). The concentration of AuPEG is calculated under the assumption that no material loss occurs during the centrifugation process.      . This value is much longer than the contour

Figures and Tables
length of a 6000 g/mol PEG chain (38 nm, calculated from 0.28 nm for the contour length of each monomer unit 4 .