Silicon carbide dust is forged in the atmospheres of dying, carbon-rich stars and later scattered across the cosmos. But exactly how these grains form has long remained a mystery. Now researchers have re-created the process in the laboratory for the first time, building silicon carbide dust analogs from atomic precursors and identifying SiC2 as a key molecular intermediate (Nat. Astron. 2026, DOI: 10.1038/s41550-026-02854-1).
“Mimicking the conditions of evolved stars is challenging. We had to develop specific instrumentation for this purpose,” says Gonzalo Santoro, one of the study’s authors who, along with members of an interdisciplinary team, spent more than 5 years designing the experiment. The effort culminated in Stardust, a custom-built, ultra-high-vacuum instrument featuring specially designed components.
Silicon carbide grains were first identified around carbon-rich asymptotic giant branch (AGB) stars roughly 50 years ago, but the chemical pathways linking gas-phase molecules to solid dust particles have remained unclear.
“In many ways, AGB stars are the ‘smokestacks’ of the galaxy,” says Xander Tielens, an astronomer at Leiden University who was not involved in the study. “They indiscriminately pollute the Milky Way with small dust particles.” Over time, he explains, this stardust becomes incorporated into new generations of stars and planets.
In the experiments, the team combined atomic carbon, atomic silicon, and molecular hydrogen under conditions mimicking the inner envelopes of carbon-rich evolved stars. Hydrogen was included because SiC2 has so far been detected only in carbon-rich stellar environments.
“With the Stardust machine and complementary ex-situ analysis, we have uniquely studied both the molecular and nanoparticle composition resulting from the chemistry of silicon and carbon atoms,” says one of the study authors, Christine Joblin of the University of Toulouse.
The results point to a bottom-up mechanism in which hydrogen first reacts with atomic carbon to generate hydrocarbons, which subsequently combine with silicon atoms to form gas-phase SiC2. The molecule then acts as a key building block for silicon carbide dust formation. Spectroscopy analysis confirms the formation of partially hydrogenated silicon carbide nanodust, and thermochemical and kinetic modeling support the proposed pathway.
“The evidence appears quite strong,” says Michael McCarthy, the deputy director at the Center for Astrophysics who was not involved in the study. “SiC2 was identified through multiple techniques. Competing candidates (Si2, SiC, Si2C, Si2C2) were searched for and not detected, and the kinetics model supports a relatively simple chemical pathway.”
“The major advantage of our machine is that we can reproduce dust formation starting from atomic precursors,” Santoro says. “This allowed us to investigate the gas-phase species that become incorporated” into nanodust.
The findings also highlight the importance of hydrogen-assisted hydrocarbon chemistry in the experimental pathway leading to SiC2, Tielens says.
“Molecular hydrogen is often treated as an inert bath gas, but in this instance, it appears to facilitate SiC2 formation by first generating simple hydrocarbons, which then subsequently react with atomic silicon,” McCarthy says. “To my knowledge, current dust-formation models do not appear to capture this coupling between the silicon chemistry and the H2-driven hydrocarbon network, so this is an important development.”
Beyond astrochemistry, the study may also offer a model system for understanding nucleation and growth processes in silicon-carbon materials more broadly. The methodology can be extended to other elements, including magnesium, aluminum, and iron, “and to the environments of oxygen-rich evolved stars,” says study coauthor José Cernicharo of the Fundamental Physics Institute.
Even so, the researchers stress that Stardust remains an approximation of real stellar environments. “There is no way to fully reproduce in the laboratory the physical and chemical conditions of these complex atmospheres,” Cernicharo says. Although the setup reproduces astrophysically relevant pressures, it cannot capture the full chemical complexity or long reaction timescales.
