Shape-Shifting Gold Clusters: How a Sol-Gel Recipe Turns Nanocatalysts Into Films That Work 2.5 Times Faster

Gold has dazzled humanity for millennia, but its most astonishing form may be the one too small to see: clusters of exactly twenty-five gold atoms, each shielded by a molecular jacket of thiol ligands. These atomically precise particles—known as Au25(SR)18 nanoclusters—behave less like bulk metal and more like giant molecules, with discrete energy levels, distinctive optical fingerprints, and catalytic talents that larger gold particles cannot match. Now, chemists at the State University of New York at Oswego have shown that these tiny gold clusters can be woven together into larger superstructures using a century-old trick of materials chemistry, and that the shape of the resulting assembly can make a dramatic difference in how well it catalyzes a reaction.

In a study published in Frontiers in Chemistry, Mahdi Hesari and Marysa Colvin describe a one-pot synthesis in which Au25 nanoclusters are locked into a silica network through a sol-gel process, producing either chunky particles or ultrathin films depending only on how the material is purified. The choice matters enormously: when tested as catalysts for the hydrogenation of 4-nitrophenol, a benchmark reaction in nanoscience, the film version ran roughly 2.5 times faster than the particle version, while both could be recovered and reused with 96 percent recovery over ten cycles. The finding suggests that for nanocluster catalysts, architecture is not a cosmetic detail—it is a performance dial.

The key to the whole scheme is a remarkably clever molecule called 3-mercaptopropyl trimethoxysilane, or MPTS. It is a dual-functional ligand: on one end sits a thiol group that bonds covalently to gold atoms, exactly like the thiols traditionally used to stabilize gold nanoclusters; on the other end sits a trimethoxysilane group, the workhorse of sol-gel chemistry. Once the silane ends are exposed to traces of water, they hydrolyze and condense, knitting together a network of Si-O-Si bonds—essentially glass—that physically connects every nanocluster to its neighbors. In other words, the same ligand that stabilizes each cluster individually also serves as the molecular glue that assembles them collectively.

The synthesis itself follows a modified one-pot route. Tetrachloroaurate is dissolved in tetrahydrofuran with tetra-n-octylammonium bromide, chilled in an ice bath, and treated with MPTS, which displaces the chloride ligands and forms Au-S bonds. A rapid injection of ice-cold sodium borohydride then reduces the gold, turning the solution a dark brown—the telltale sign of nanocluster formation. After 24 hours of stirring, gravity filtration removes unreacted gold-thiolate polymer intermediates, leaving a bright brown solution ready for the sol-gel stage. From there, the two purification pathways diverge, and it is that divergence that decides the final shape of the material.

In procedure A, the crude product is rotary-evaporated to an oily residue and washed with ethanol, the conventional purification for these clusters. In the presence of residual water in the THF solution and added methanol, the silane groups condense into a solid sol-gel network, yielding the Au25-SiO2 particle form—dark brownish chunks whose individual particles average roughly 62 by 49 micrometers. Procedure B takes a gentler route: size exclusion chromatography on a Sephadex LH-20 column separates the clusters from excess thiol based on molecular weight, and the solvent is removed by rotary evaporation at room temperature. Because the excess thiol is stripped away, no ethanol or methanol is introduced, and the water content stays low, the clusters instead condense into a continuous film just 12.76 nanometers thick—roughly a thousand times thinner than the particles are wide.

Confirming that the delicate Au25 clusters survive both the gelation and the solidification required a battery of analytical techniques. Solid-state UV-visible spectroscopy revealed the cluster’s signature features, including absorption maxima and valleys matching those reported for Au25(SR)18 species in solution, in both the particle and the film. Mass spectrometry of the soluble precursor detected multiply charged fragments of the intact cluster, along with telltale losses of methyl groups from the ligand shells—51 such losses observed against a theoretical count of about 54. X-ray photoelectron spectroscopy mapped the elemental landscape of the solids, detecting gold, silicon, sulfur, carbon, nitrogen, and oxygen, with the nitrogen 1s signal at 401.96 electron-volts confirming the tetra-octylammonium counterion and thus the negative charge state of the embedded clusters. Scanning electron microscopy with energy-dispersive X-ray analysis showed gold, silicon, and sulfur distributed homogeneously across both forms, and X-ray diffraction revealed broad peaks consistent with compact face-centered-cubic gold domains and the amorphous silica texture of the gel.

Perhaps the most reassuring result came from photoluminescence. When the film was excited with a 532-nanometer laser, it emitted near-infrared light with a peak near 940 nanometers and a shoulder at 765 nanometers—essentially the same emission spectrum reported for purified Au25 clusters in solution. Those emissions arise from intraband transitions between the cluster’s quantized electronic states, so their persistence in the solid state is strong evidence that the gold cores remain intact and electronically unperturbed inside the silica matrix. The team even observed a partial enhancement of the luminescence, which they attribute to energy transfer between neighboring embedded clusters—a hint that the assembled superstructure is more than the sum of its parts, precisely as the superstructure concept predicts.

The catalytic test case was the reduction of 4-nitrophenol to 4-aminophenol by sodium borohydride, a fast, well-studied reaction that serves as a standard yardstick for metal nanocatalysts. Monitoring the decay of the 4-nitrophenol absorption at 400 nanometers, the researchers found that after 15 minutes the film had converted 92 percent of the substrate while the particle had managed 81 percent. The extracted rate constants were 2.44 × 10⁻³ per second for the film and 9.83 × 10⁻⁴ per second for the particle—a 2.5-fold advantage for the film. The explanation is structural: the film’s silica layer is far thinner, giving reactants more effective access to the gold surfaces, whereas the thicker silica coating of the particles imposes a longer diffusion path. The results also demonstrate that the silica content is calibrated just right—enough to weld the clusters into a robust network, but not so much that it buries the active sites.

For practical catalysis, the silica shell delivers a crucial bonus: recyclability. Bare Au25 clusters in solution are fast catalysts—some reported systems run considerably faster than the silica-embedded versions—but they are difficult to recover and can degrade at the elevated temperatures and harsh chemical conditions that heterogeneous catalysis often demands. The Au25-SiO2 solids, by contrast, can simply be filtered out of the reaction mixture and used again, retaining their activity over at least ten cycles with 96 percent recovery. The authors note that the measured rates sit close to published values for other silica-coated and water-soluble gold and silver-copper nanocluster catalysts, placing this material squarely within the competitive landscape of modern nanocatalysis while adding the durability that industrial applications require.

Beyond the immediate catalytic numbers, the study opens a general recipe: any atomically precise metal nanocluster protected with a silane-functionalized thiol could, in principle, be fused into particle, film, or potentially other morphologies simply by tuning the purification conditions. The authors suggest that synchrotron spectroscopy could further illuminate how structure and performance are linked in these morphologies, and that the same assembly strategy might extend across the rich family of Au, Ag, and alloy nanoclusters now being synthesized worldwide. What began as a routine purification step has turned into a design principle—proof that when it comes to atomically precise catalysts, how you wash your chemistry can determine how it works.

Subject of Research: Sol-gel assembly of Au25 gold nanoclusters into silica particles and films for enhanced catalysis

Article Title: Controlled sol-gel nanoassembly of Au25(SR)18 nanoclusters for structural and functional integration

Article References: Controlled sol-gel nanoassembly of Au25(SR)18 nanoclusters for structural and functional integration. (n.d.). https://doi.org/10.3389/fchem.2026.1982602

Image Credits: AI Generated

DOI: 10.3389/fchem.2026.1982602

Keywords: gold nanoclusters, Au25, sol-gel, nanoassembly, silica, catalysis, nitrophenol hydrogenation, MPTS, thin films, heterogeneous catalysis, photoluminescence, XPS