Unevolved Li-rich stars at low metallicity: a possible formation pathway through novae

Abstract: A small fraction of low-mass stars have been found to have anomalously high Li abundances. Although it has been suggested that mixing during the red giant branch phase can lead to Li production, this method of intrinsic Li production cannot explain Li-rich stars that have not yet undergone the first dredge-up. To obtain clues about the origin of such stars, we present a detailed chemical abundance analysis of four unevolved Li-rich stars with $-2.1 < [\mathrm{Fe/H}] < -1.3$ and $2.9<A({\rm Li})<3.6$, $0.7-1.4$ dex higher Li abundance than typical unevolved metal-poor stars. One of the stars, Gaia DR3 6334970766103389824 (D25_6334), was serendipitously found in the stellar stream ED-3, and the other three stars have been reported to have massive ($M\gtrsim 1.3\,\mathrm{M_\odot}$) non-luminous companions. We show that three of the four stars exhibit abundance patterns similar to those of known unevolved Li-rich stars, namely normal abundances in most elements except for Li and Na. These abundance similarities suggest a common origin for the unevolved Li-rich stars and low-mass metal-poor stars with massive compact companions. We also made the first detection of N abundance to unevolved Li-rich stars in D25_6334, and found that it is significantly enhanced ($[\mathrm{N/Fe}]=1.3$). The observed abundance pattern of D25_6334, spanning from C to Si, indicates that its surface has been polluted by an intermediate-mass former companion star or a nova system that involves a massive ONe white dwarf. Using a population synthesis model, we show that the nova scenario can lead to the observed level of Li enhancement and also provide an explanation for Li-rich stars without companions and those with massive compact companions.