Fossil imprints of the first-generation supernova ejecta in extremely metal-deficient stars

Using the results of nucleosynthesis calculations for theoretical con-colla pse supernova models with various progenitor masses, it is shown that the a bundance patterns of C, Mg, Si, Ca, and H that are seen in extremely metal- deficient stars with [Fe/H] less than or similar to -2.5 follow those seen in the individual first-generation supernova remnants (SNRs). This suggests that most of the stars with [Fe/H] less than or similar to -2.5 were made from individual supernova (SN) events. To obtain the ratio of heavy element s to hydrogen, a formula is derived to estimate the mass of hydrogen swept up by an SNR when it occurs in the interstellar matter with the primordial abundances. We use [Mg/H] to indicate the metallicities instead of [Fe/H]. The metallicities [Mg/H] predicted from these SNRs range from similar to-4 to similar to-1.5, and the mass of Mg in an SN is well correlated with its progenitor mass. Thus, the observed [Mg/H] in an extremely metal-deficient star has a correspondence to the progenitor mass. A larger [Mg/H] correspon ds to a larger progenitor mass. Therefore, the so-called "age-metallicity r elation" does not hold for stars with [Fe/H] less than or similar to -2.5. In contrast, the [Mg/Fe] ratios in the theoretical SNRs have a different tr end from those in extremely metal-deficient stars. It is also shown that fr om the observed trend of [Mg/Fe], one can predict the Fe yield of each SN g iven the correspondence of [Mg/H] to the progenitor mass. The Fe yields thu s obtained are consistent with those derived from SN light-curve analyses. This indicates that there is still a problem in modeling a core-collapse su pernova at the beginning of its explosion or mass cut. The abundance determ ination of O in extremely metal-deficient stars, which has not been done fr om observational analyses, is strongly desired in order to test the hypothe sis that the elements in an extremely metal-deficient star come from a sing le SN event and to obtain reliable yields for SNe.