Exploring the ultra-faint dwarf Bootes I using JWST and HST: Metallicity distribution and binaries

TL;DR

This paper tackles how to recover the metallicity distribution and binary content of the ultra-faint dwarf Boötes I by combining deep HST and JWST photometry that extends to M-dwarfs. The authors introduce a Metallicty inference via faint MS colors and a Binary Map to constrain the MS binary population, applying artificial-star simulations to calibrate completeness and mass-ratio effects. They find ⟨[Fe/H]⟩ = $-2.53 \pm 0.05$ with $\sigma_{[Fe/H]} = 0.41 \pm 0.03$ dex, and a total binary fraction of $f_{bin}^{TOT} = 0.30 \pm 0.03$ with $f_{bin}^{q>0.4} = 0.20 \pm 0.02$, implying a substantial binary presence similar to open and Magellanic Cloud clusters. The results suggest that dark matter does not strongly modify binary properties in Boötes I and demonstrate a powerful photometric path to metallicity and binary-characterization in UFDs, enabling broader applications to faint stellar systems where RGB spectroscopy is challenging.

Abstract

Ultra-faint dwarf galaxies (UFDs) are among the oldest and most metal-poor stellar systems in the Universe. Their metallicity distribution encodes the fossil record of the earliest star formation, feedback, and chemical enrichment, providing crucial tests of models of the first stars, galaxy assembly, and dark matter halos. However, due to their faint luminosities and the limited number of bright giants, spectroscopic studies of UFDs typically probe only small stellar samples. Here, we present an analysis of multi-epoch Hubble Space Telescope and James Webb Space Telescope observations of the UFD Bootes I. Using deep color-magnitude diagram in the F606W and F322W2 bands, extending from the subgiant branch to the M-dwarfs, and stellar proper motions to identify likely members, we obtained an unprecedentedly clean census of the system. The exquisite quality of the diagram, combined with the sensitivity of M-dwarf colors to metallicity, allowed us to constrain the metallicity distribution in a large stellar sample. As a first step, we derived the binary fraction in Bootes I. This is crucial, since binaries can bias kinematic mass estimates, affect stellar population analyses, and shape the photometric signatures used to infer metallicity. We find that 20$\pm$2% of stellar systems in Bootes I are binaries with mass ratios larger than 0.4, corresponding to a total binary fraction of $\sim$30%. This value is comparable to the binary fractions observed in globular clusters of similar stellar mass, suggesting that the presence of dark matter does not significantly affect the binary properties of Bootes I. We then exploited the metallicity sensitivity of M-dwarf colors to derive the metallicity distribution function. We find that most of the stars $\sim$85% have [Fe/H]<-2, and that roughly $\sim$17% have [Fe/H]<-3.

Figures (10)

• Figure 1: Footprints of the HST and JWST images used in this work are shown in red and green, respectively, in the top-left panel. The bottom-left panel displays the stacked NIRCam/F322W2 image used in our analysis. The right panel shows a three-colour composite image of the region marked with a red square in the bottom-left panel, where the blue, green, and red channels correspond to the stacked F606W, F814W, and F322W2 images, respectively.
• Figure 2: $m_{\rm F322W2}$ vs. $m_{\rm F606W}-m_{\rm F322W2}$ CMD of Boötes I. The red bars indicate the average photometric errors.
• Figure 3: $m_{\rm F322W2}$ vs. $m_{\rm F606W}-m_{\rm F322W2}$ CMD of Boötes I (left panel) and $m_{\rm F322W2}$ vs. $\Delta_{\rm F606W,F322W2}$ verticalized diagram of the lower MS (right panel). We superimpose on the CMD nine isochrones with varying [Fe/H] and [$\alpha$/Fe], as indicated in the inset. The isochrones used to verticalize the diagram are shown with continuous lines.
• Figure 4: Kernel (top) and cumulative (bottom) distributions of $\Delta_{\rm F606W,F322W2}$ for faint MS stars. The black curves are derived from the observed Boötes I members, while the coloured curves are obtained from the simulated CMDs.
• Figure 5: Left: Comparison between the observed (black) and best-fit simulated (green) kernel (top) and cumulative (bottom) distributions of $\Delta_{\rm F606W,F322W2}$. Middle: [Fe/H] histogram that best reproduces the observations of Boötes I. The dashed black distribution is from a spectroscopic survey from lai2011. Right: Comparison between the observed (top) and simulated (bottom) CMDs, zoomed in on the MS region used to infer the metallicity distribution. The colors of the simulated stars indicate their iron abundance, as shown in the middle panel.