Evolutionary link between globular clusters and circumgalactic clouds

The established by us possibility to consider circumgalactic clouds (CGCs) as the remnants of the parent clouds in which globular clusters (GCs) have been formed (Acharova & Sharina 2018) is based on a comparison of the following facts. First, the metallicities of CGCs at redshifts $z<1$ and of GCs in our and other galaxies show a bimodal distribution with a minimum near $\rm [Mg/H]=-1$. Mean values and standard deviations of the Mg abundances in GCs and CGCs with $[\mathrm{M}\mathrm{g}\mathrm{/}\mathrm{H}]<-1$ and $[\mathrm{M}\mathrm{g}\mathrm{/}\mathrm{H}]>-1$ coincide within the typical error of measuring the elemental abundances in clouds: 0.3 dex (Acharova & Sharina 2018). Second, a similar coincidence is observed for GCs and CGCs with $[\mathrm{X}\mathrm{/}\mathrm{H}]<-1$ and $[\mathrm{X}\mathrm{/}\mathrm{H}]>-1$ at redshifts $2<z<3$, where $[X\mathrm{/}H]$ is the metallicity determined from the sum of several elemental abundances (Dias et al. 2016, Rafelski et al. 2012, Wotta et al. 2019, Quiret et al. 2016). Third, high-metallicity CGCs are observed starting from redshifts $\rm z\le 2.5$, i.e. approximately 11 Gyrs ago. At the same time globular clusters were actively formed, and their supernovae were able to enrich the surrounding gas, from which the high-metal component of the clouds was formed.