Supernovae ( SNe ) are considered to have a major role in dust enrichment of high redshift galaxies and , due to the short lifetimes of interstellar grains , in dust replenishment of local galaxies .
Here we explore how SN dust yields depend on the mass , metallicity , and rotation rate of the progenitor stars , and on the properties of the explosion .
To this aim , assuming uniform mixing inside the ejecta , we quantify the dust mass produced by a sample of SN models with progenitor masses 13 ~ { } M _ { \odot } \leq M \leq 120 ~ { } M _ { \odot } , metallicity -3 \leq [ Fe / H ] \leq 0 , rotation rate v _ { rot } = 0 and 300 km/s , that explode with a fixed energy of 1.2 \times 10 ^ { 51 } erg ( FE models ) or with explosion properties calibrated to reproduce the ^ { 56 } Ni - M relation inferred from SN observations ( CE models ) .
We find that rotation favours more efficient dust production , particularly for more massive , low metallicity stars , but that metallicity and explosion properties have the largest effects on the dust mass and its composition .
In FE models , SNe with M \leq 20 - 25 ~ { } M _ { \odot } are more efficient at forming dust : between 0.1 and 1 M _ { \odot } is formed in a single explosion , with a composition dominated by silicates , carbon and magnetite grains when [ Fe / H ] = 0 , and by carbon and magnetite grains when [ Fe / H ] < 0 .
In CE models , the ejecta are massive and metal-rich and dust production is more efficient .
The dust mass increases with M and it is dominated by silicates , at all [ Fe/H ] .