Table of Contents
Fetching ...

The origin of strong $α$-element bimodalities in FIRE simulations of Milky Way-mass galaxies

Megan Barry, Andrew Wetzel, Sarah Loebman, Jeremy Bailin, Hanna Parul

TL;DR

This study analyzes the origin of strong $\alpha$-element bimodality in FIRE-2 Milky Way–mass galaxies by examining $[\mathrm{Mg}/\mathrm{Fe}]$ vs $[\mathrm{Fe}/\mathrm{H}]$ across 16 systems. Four galaxies exhibit a pronounced bimodality, with high-$\alpha$ stars in old, radially compact thick disks and low-$\alpha$ stars in younger, radially extended thin disks, paralleling the Milky Way. The bimodality arises during a relatively brief transition (0.3–1.2 Gyr) that follows a sharp drop in star formation and occurs in a low-gas-fraction phase; Fe enrichment by white-dwarf SNe rises relative to Mg from core-collapse SNe, driving $[\mathrm{Mg}/\mathrm{Fe}]$ downward at nearly fixed $[\mathrm{Fe}/\mathrm{H}]$. Importantly, the formation of the bimodality does not require a major merger or significant radial migration, though a radial $[\mathrm{Fe}/\mathrm{H}]$ gradient tends to form during the transition, suggesting a condition that facilitates bimodality. Overall, strong $\alpha$ bimodalities in MW-mass galaxies are relatively rare and reflect complex, largely epoch-dependent SFH and gas-accretion histories rather than a single universal trigger.

Abstract

One of the Milky Way's characteristic features is a strongly bimodal distribution of $α$-process elements, such as Mg, at fixed [Fe/H] in stellar abundances. We examine patterns in [Mg/Fe] versus [Fe/H] in FIRE-2 simulations of Milky Way-mass galaxies. Out of 16 galaxies, 4 are capable of producing a strongly bimodal distribution. In all four galaxies, the high-$α$ population corresponds to an older, radially-compact, thick disk, and the low-$α$ population corresponds to a younger, radially-extended, thin disk, similar to the MW.The transition from high- to low-$α$ took $0.3-1.2\Gyr$ and began $5.5-6.5\Gyr$ ago. [Mg/Fe] decreased at relatively fixed [Fe/H], both in the galaxy overall and at fixed radii: Fe enrichment nearly balanced gas accretion (and therefore dilution), but Mg enrichment was weaker. Importantly, this transition occurred during a period of relatively low gas fraction ($5-15\%$), immediately after a rapid decline in star formation (halving within a few hundred Myr), which caused an increase in Fe-producing white-dwarf supernovae relative to Mg-producing core-collapse supernovae. Only one case coincided with a major merger coalescence. We find similar trends in measuring stars by their current radius and by their birth radius, therefore, radial redistribution did not play a dominant role in the formation of a bimodality or its spatial dependence today. Overall, in FIRE-2, strong $α$-element bimodalities are relatively uncommon ($\sim25\%$), often not associated with a major merger, and arise primarily from a rapid decline in star formation during relatively low gas fraction.

The origin of strong $α$-element bimodalities in FIRE simulations of Milky Way-mass galaxies

TL;DR

This study analyzes the origin of strong -element bimodality in FIRE-2 Milky Way–mass galaxies by examining vs across 16 systems. Four galaxies exhibit a pronounced bimodality, with high- stars in old, radially compact thick disks and low- stars in younger, radially extended thin disks, paralleling the Milky Way. The bimodality arises during a relatively brief transition (0.3–1.2 Gyr) that follows a sharp drop in star formation and occurs in a low-gas-fraction phase; Fe enrichment by white-dwarf SNe rises relative to Mg from core-collapse SNe, driving downward at nearly fixed . Importantly, the formation of the bimodality does not require a major merger or significant radial migration, though a radial gradient tends to form during the transition, suggesting a condition that facilitates bimodality. Overall, strong bimodalities in MW-mass galaxies are relatively rare and reflect complex, largely epoch-dependent SFH and gas-accretion histories rather than a single universal trigger.

Abstract

One of the Milky Way's characteristic features is a strongly bimodal distribution of -process elements, such as Mg, at fixed [Fe/H] in stellar abundances. We examine patterns in [Mg/Fe] versus [Fe/H] in FIRE-2 simulations of Milky Way-mass galaxies. Out of 16 galaxies, 4 are capable of producing a strongly bimodal distribution. In all four galaxies, the high- population corresponds to an older, radially-compact, thick disk, and the low- population corresponds to a younger, radially-extended, thin disk, similar to the MW.The transition from high- to low- took and began ago. [Mg/Fe] decreased at relatively fixed [Fe/H], both in the galaxy overall and at fixed radii: Fe enrichment nearly balanced gas accretion (and therefore dilution), but Mg enrichment was weaker. Importantly, this transition occurred during a period of relatively low gas fraction (), immediately after a rapid decline in star formation (halving within a few hundred Myr), which caused an increase in Fe-producing white-dwarf supernovae relative to Mg-producing core-collapse supernovae. Only one case coincided with a major merger coalescence. We find similar trends in measuring stars by their current radius and by their birth radius, therefore, radial redistribution did not play a dominant role in the formation of a bimodality or its spatial dependence today. Overall, in FIRE-2, strong -element bimodalities are relatively uncommon (), often not associated with a major merger, and arise primarily from a rapid decline in star formation during relatively low gas fraction.
Paper Structure (18 sections, 1 equation, 10 figures)

This paper contains 18 sections, 1 equation, 10 figures.

Figures (10)

  • Figure 1: Left: 2-D histogram of [Mg/Fe] versus [Fe/H] today, color-coded linearly by the mass in stars, for the 4 MW-mass galaxies in the FIRE-2 simulations with a strong bimodality in $\alpha$-elements. The oldest stars populate the upper-left region, while the youngest stars populate the lower region. Each yellow curve delineates the boundary between well-defined high- and low-$\alpha$ populations (see Section \ref{['subsec:define']}). We include stars to the left of this region as high-$\alpha$ and stars to the right of this region as low-$\alpha$. Right: Images of stars today, color-coded by logarithmic mass, separately for the high- and low-$\alpha$ populations. All 4 galaxies are similar to the MW, in terms of the spatial correlations of high-$\alpha$ stars with the older, radially compact, thick disk (average age $t_{\rm birth} = 8.5 \Gyr$, average $R_{\rm star,95} = 7.5 \kpc$, and average $Z_{\rm star,95} = 1.6 \kpc$), and of low-$\alpha$ stars with the younger, radially extended, thin disk (average age $t_{\rm birth} = 3.1 \Gyr$, average $R_{\rm star,95} = 10.3 \kpc$, and average $Z_{\rm star,95} = 1.0 \kpc$).
  • Figure 2: [Mg/Fe] versus [Fe/H] of stars at different spatial locations ($R$ and $Z$) in the 4 FIRE-2 galaxies with a strong bimodality in $\alpha$-elements, based on stellar positions today (left column) and at birth (right column). Both columns are measured in reference to $R_{\rm star,95}$ measured today. The trends for stars today qualitatively agree with the MW Anders_2014Nidever_2014Hayden_2015Vincenzo_2021, in terms of low-$\alpha$ stars dominating at small $Z$ and large $R$, and high-$\alpha$ stars dominating at large $Z$ and small $R$. Similar to the MW, low-$\alpha$ stars have a negative radial gradient in [Fe/H]. The main exception is m12r, the lowest-mass galaxy, which shows significant high- and low-$\alpha$ populations across the entire disk. As the right column shows, these trends for stars were largely present at birth, so they were not caused by radial redistribution of stars after birth. The main difference is that at birth there were fewer stars at high $Z$, which arises from dynamical heating of stars over time.
  • Figure 3: Left: Histogram of [Mg/Fe] versus [Fe/H] today, as in Figure \ref{['fig:disks']}. Each white curve shows the path over time (beginning at the top left) of the median [Mg/Fe]-[Fe/H] of newly-formed stars across the galaxy. Colored dots show the beginning ($t_{\rm onset}$), middle ($t_{\rm mid}$), and end ($t_{\rm end}$) of the high-to-low $\alpha$ transition, as defined in Section \ref{['sec:motiv']} and listed in Table \ref{['tab:galaxies']}. Most galaxies underwent the $\alpha$ transition, evolving from high- to low-[Mg/Fe], at relatively fixed [Fe/H], following a downward "turn" in the evolutionary track.Right: Top panels shows the low-$\alpha$ fraction, $f_{\rm low-\alpha} = M_{\rm star, low-\alpha} / M_{\rm star, tot}$, versus stellar age. Vertical dashed lines indicate key times of the $\alpha$ transition: $t_{\rm onset}$, $t_{\rm mid}$, $t_{\rm end}$. Shaded regions show the duration of the $\alpha$ transition, which ranges from $0.3 \Gyr$ (m12i) to $1.2 \Gyr$ (m12r). We show 2 instances of $t_{\rm onset}$ for m12q (bottom), because $f_{\rm low-\alpha}$ rises but then falls to $\lesssim 0.2$ for nearly 1 Gyr before beginning to increase again. Bottom panels show average [Mg/H] and [Fe/H] of stars at birth versus age. We normalize [Fe/H] so [Mg/H] and [Fe/H] are equal $0.5 \Gyr$ before $t_{\rm onset}$ to compare their evolution and divergence. During the $\alpha$ transition, [Fe/H] remained relatively fixed, via a near-balance between (strong) Fe enrichment and gas accretion (dilution), while [Mg/H] generally declined, because of less rapid Mg enrichment.
  • Figure 4: Key physical properties versus lookback time that drove the strong $\alpha$-element bimodality in these 4 galaxies. Vertical dashed lines and gray shaded regions show the $\alpha$ transition period, as defined by $f_{\rm low-\alpha}$ in Figure \ref{['fig:alphafrac']}. Top rows: Each red curve shows the star formation rate (SFR): the thick curve shows SFR smoothed over 50 Myr, highlighting a sharp drop (up to a twofold decrease) in SFR immediately before the $\alpha$ transition, while the thin curve shows SFR in bins of 10 Myr, which shows a transition from bursty to smooth star formation when the SFR declined for all cases except m12r. Each blue curve shows the instantaneous ratio of the (Fe-producing) white-dwarf supernova (WDSN) rate to the ($\alpha$-producing) core-collapse supernova (CCSN) rate, which rose following the drop in SFR, typically approximately doubling. In all 4 galaxies, the $\alpha$ transition occurred after or during a drop in SFR, corresponding to an increase in the WD/CC SN rate, and thus more enrichment in Fe than Mg. Bottom rows: Gas fraction, $f_{\rm gas} = M_{\rm gas} / (M_{\rm gas} + M_{\rm star})$, calculated within $R_{\rm star,95}$ and $Z_{\rm star,95}$ at that snapshot. The $\alpha$ transition occurred after a $\approx 50\%$ decrease in $f_{\rm gas}$ for m12r, m12i, and m12q. In all 4 cases, $f_{\rm gas}$ remained nearly constant across the transition period, despite ongoing star formation, indicating ongoing gas accretion and dilution.
  • Figure 5: Paths through [Mg/Fe]-[Fe/H], as in Figure \ref{['fig:alphafrac']}, now split into bins of galactocentric radii, for both $R_{\rm star,now}$ (top) and $R_{\rm star,now}$ (bottom), in units of $R_{\rm star,95}$. Both show similar trends, further highlighting that the key patterns of the $\alpha$ bimodality were present for stars at birth and were not significantly affected by radial redistribution of stars after birth. The erratic behavior in m12i arises from bins of low star formation. Overall, stars at different $R$ had similar abundances during the high-$\alpha$ period, but stars in different $R$ split apart around the time of the transition to the low-$\alpha$ phase, because this coincided with the formation of a radial gradient in metallicity. Even at fixed $R$, stars evolved from high- to low-[Mg/Fe] during the $\alpha$ transition period at relatively fixed [Fe/H].
  • ...and 5 more figures