CosmicGrain follows dust’s individual histories through a simulated galaxy

Cosmic dust can store the heavy elements needed for future stars, exchange heat with gas and respond to starlight. Yet following its evolution inside a growing galaxy requires connecting tiny grains with environments spanning enormous distances. A new simulation module, CosmicGrain, tackles that problem by giving dust its own motion and evolving history rather than simply carrying it along with gas.

In a preprint posted on November 4, Ezra Huscher and Kristian Finlator describe the module and experiments following one galaxy and its surrounding halo through cosmic time. Their results suggest that the histories of individual dust populations matter for both their survival and their eventual location. They also expose substantial limitations in the current implementation.

Giving dust a separate trajectory

Built on the Gadget-4 simulation code, CosmicGrain represents dust using “superparticles.” Each stands for a population of physical grains with one representative radius, not a single microscopic speck. That radius changes as material accumulates, erodes or fragments, while the superparticle moves under gravity, gas drag and radiation pressure.

This allows different dust populations to encounter different temperatures and densities. A grain population passing through hot gas can suffer a different fate from one sheltered in a denser region, even if both began with similar properties.

The researchers tested an eleven-step physics ladder. Starting with dust creation alone, they progressively added thermal exchange, drag, incorporation into forming stars, erosion by hot gas, growth, a correction for unresolved dense clouds, supernova-shock destruction, sticking collisions, fragmentation and radiation pressure. Because each step builds on the previous one, its effect describes that particular sequence—not a universal contribution independent of the other processes.

Drag keeps dust nearby

Gas drag produces the ladder’s most dramatic change in where dust resides. Without it, dust travels outward from its stellar sources and little remains in the inner galaxy. Adding drag raises the inner region’s dust-to-gas ratio by more than two orders of magnitude, while increasing the dust-to-stellar mass ratio across the halo by roughly threefold.

The distinction matters: retaining dust near the galaxy is not the same as creating more dust. Drag slows grains relative to nearby gas, helping prevent their escape from the central region. Larger grains have longer stopping times, allowing them to travel farther before their motion adjusts to the gas.

Hot-gas erosion, known as thermal sputtering, then sharply reduces the surviving population. Grain growth replenishes some material, particularly after the model boosts effective gas densities to represent cold clouds that the simulation cannot resolve. That boost is a numerical prescription, not a direct simulation of molecular-cloud structure.

Birth size leaves a lasting mark

One striking result concerns fragmentation. In the shattering experiment, the authors report 38,018 complete destructions of silicate-dominated superparticles, compared with just 17 carbon-dominated ones.

Those numbers depend on the model’s starting assumptions. Supernova-sourced, silicate-rich grains begin with radii of 10 nanometres, while carbon-rich grains from aging stars begin at 100 nanometres. Fragmentation repeatedly reduces grain radius, so the initially smaller population needs fewer events to fall below the model’s destruction threshold. Carbon-rich grains also have a higher fragmentation-velocity threshold.

The result illustrates how composition and inherited size can work together, rather than establishing a universal destruction ratio for real galaxies. The full model also preferentially places carbon-rich dust in the surrounding circumgalactic medium. Radiation pressure helps redistribute larger grains outward, although its strength is calibrated using an observed dust-profile slope.

A useful experiment, not a finished census

The authors find broadly reasonable agreement with some observed radial dust ratios and dust-to-metal measurements. But their galaxy is dust-rich compared with typical observed systems at similar stellar mass, and the paper identifies a major missing destruction effect.

Supernova-shock destruction is effectively negligible in these runs. The physical blast radius is around 50 parsecs, much smaller than the 3,000-parsec search region used to find dust. A volume correction suppresses destruction too strongly, the authors explain. Its weak effect therefore cannot establish that supernova shocks are unimportant in nature.

Resolution remains another caution. The draft’s convergence discussion does not consistently match its table: the tabulated halo-wide dust-to-gas ratio rises from 0.0011 at the main resolution to 0.012 at the highest resolution, rather than declining as the prose states.

Together with the single-halo focus, simplified radiation treatment and absence of dust’s drag back-reaction on gas, these limitations restrict quantitative predictions. CosmicGrain’s clearest contribution is an openly available framework for investigating how dust’s trajectories and accumulated processing shape a galaxy’s dust population.