Signatures of very massive stars: supercollapsars and their cosmological rate

We compute the rate of supercollapsars by using cosmological, N-body, hydro, chemistry simulations of structure formation, following detailed stellar evolution according to proper yields (for He, C, N, O, Si, S, Fe, Mg, Ca, Ne, etc.) and lifetimes for stars having different masses and metallicities, and for different stellar populations (population III and population II-I). We find that supercollapsars are usually associated to dense, collapsing gas with little metal pollution and with abundances dominated by oxygen. The resulting supercollapsar rate is about $10^{-2}\,\rm yr^{-1} sr^{-1}$ at redshift $z=0$, and their contribution to the total rate is $ < 0.1 $ per cent, which explains why they have never been detected so far. Expected rates at redshift $z\simeq 6$ are of the order of $\sim 10^{-3}\,\rm yr^{-1} sr^{-1}$ and decrease further at higher $z$. Because of the strong metal enrichment by massive, short-lived stars, only $\sim 1$ supercollapsar generation is possible in the same star forming region. Given their sensitivity to the high-mass end of the primordial stellar mass function, they are suitable candidates to probe pristine population III star formation and stellar evolution at low metallicities.