The missing metals problem. III How many metals are expelled from galaxies?

Abstract

[Abridged] We revisit the metal budget at z~2. In the first two papers of this series, we already showed that ~30% (to <60% if extrapolating the LF) of the metals are observed in all z~2.5 galaxies detected in current surveys. Here, we extend our analysis to the metals outside galaxies, i.e. in intergalactic medium (IGM), using observational data and analytical calculations. Our results for the two are strikingly similar: (1) Observationally, we find that, besides the small (5%) contribution of DLAs, the forest and sub-DLAs contribute subtantially to make <30--45% of the metal budget, but neither of these appear to be sufficient to close the metal budget. The forest accounts for 15--30% depending on the UV background, and sub-DLAs for >2% to <17% depending on the ionization fraction. Together, the `missing metals' problem is substantially eased. (2) We perform analytical calculations based on the effective yield--mass relation. At z=2, we find that the method predicts that 2$--50% of the metals have been ejected from galaxies into the IGM, consistent with the observations. The metal ejection is predominantly by L<1/3L_B^*(z=2) galaxies, which are responsible for 90% the metal enrichment, while the 50 percentile is at L~1/10L^*_B(z=2). As a consequence, if indeed 50% of the metals have been ejected from galaxies, 3--5 bursts of star formation are required per galaxy prior to z=2. The ratio between the mass of metals outside galaxies to those in stars has changed from z=2 to z=0: it was 2:1 or 1:1 and is now 1:8 or 1:9. This evolution implies that a significant fraction of the IGM metals will cool and fall back into galaxies.