Gravitational Fragmentation of Extremely Metal-poor Circumstellar Discs

Abstract

We study the gravitational fragmentation of circumstellar discs accreting extremely metal-poor (𝑍 ≤ 10 −3 Z ) gas, performing a suite of three-dimensional hydrodynamic simulations using the adaptive mesh refinement code Enzo. We systematically follow the long-term evolution for 2 × 10 3 years after the first protostar's birth, for the cases of 𝑍 = 0, 10 −5 , 10 −4 , and 10 −3 Z . We show that evolution of number of self-gravitating clumps qualitatively changes with 𝑍. Vigorous fragmentation induced by dust cooling occurs in the metal-poor cases, temporarily providing ∼ 10 self-gravitating clumps at 𝑍 = 10 −5 and 10 −4 Z . However, we also show that the fragmentation is a very sporadic process; after an early episode of the fragmentation, the number of clumps continuously decreases as they merge away in these cases. The vigorous fragmentation tends to occur later with the higher 𝑍, reflecting that the dust-induced fragmentation is most efficient at the lower density. At 𝑍 = 10 −3 Z , as a result, the clump number stays smallest until the disc fragmentation starts in a late stage. We also show that the clump mass distribution also depends on the metallicity. A single or binary clump substantially more massive than the others appear only at 𝑍 = 10 −3 Z , whereas they are more evenly distributed in mass at the lower metallicities. We suggest that the disc fragmentation should provide the stellar multiple systems, but their properties drastically change with a tiny amount of metals.

Transition of the initial mass function in the metal-poor environments

“…( 4). Note that the decrease due to the mergers in this metallicity range is also reported by Shima & Hosokawa (2021), whose study follows the evolution in a few thousand years. Our result suggests that the number of stars increases in a later stage than calculated in their study and behaves similarly to the primordial case.…”

“…We have followed the stellar mass evolution until the total stellar mass reaches ∼ 150 M , at which epoch the radiation feedback will operate (see section 4.2). Although our simulation is extended longer than the previous studies (Clark et al 2008;Safranek-Shrader et al 2016;Chiaki & Yoshida 2020;Shima & Hosokawa 2021), the total stellar mass is still small compared to observed star clusters, whose masses range from ∼ 100 to 10 5 M (e.g. Krumholz et al 2019).…”

Stellar winds and metal enrichment from fast-rotating Population III stars

“…Recent advancements in hydrodynamic simulations of primordial star-forming clouds have shown that fragmentation is also present in extremely metalpoor/free environments and Pop III stars are likely formed in small clusters of a few members (see e.g. Haemmerlé et al 2020;Shima & Hosokawa 2021). In light of this, we take Pop III clusters as the basic units of Pop III formation and source of early metal enrichment, rather than individual stars.…”