Tracing ionized gas kinematics in Lyman-Break Analogs. Implications for star formation compactness and outflow properties

TL;DR

This study uses high-resolution optical spectroscopy to decompose the ionized gas in 14 local Lyman-Break Analogs into narrow, virialized and broad, outflowing components. The broad components imply ionized winds with $v_{out} \sim 200$–$500$ km s$^{-1}$ and mass-loading factors $\eta \sim 0.03$–$0.81$, with a mean around $0.31$, and with outflow rates correlating with star-formation-rate surface density. The narrow component luminosities follow the giant H II region $L$–$\sigma$ relation, while broad components indicate non-virial turbulent gas driven by stellar winds and supernovae; densities differ between components, suggesting distinct physical conditions. The results align LBAs with high-redshift star-forming systems, supporting a picture where compact starbursts drive effective feedback, and they highlight the need for spatially resolved, multiwavelength observations to map multiphase outflows across cosmic time.

Abstract

We present a study of the ionized gas kinematics and feedback properties in a sample of 14 low-mass, UV-luminous Lyman Break Analogs (LBAs) at redshifts z~0.1-0.3. These compact, strongly star-forming galaxies serve as local analogs of high-redshift starbursts. Using high-resolution VLT/X-shooter spectra, we model the optical emission-line profiles, including [O III] 4959,5007 and the Balmer lines, with multi-component Gaussian fits. All galaxies show complex kinematics that require both narrow (sigma < 90 km/s) and broad (sigma > 90 km/s) components. The narrow components trace highly turbulent gas associated with massive star-forming regions, while the broad components indicate ionized outflows driven by stellar winds and supernova feedback, with outflow velocities of about 200-500 km/s. Estimated mass outflow rates range from 0.20 to 2.72 solar masses per year, with mass-loading factors between 0.03 and 0.81. We find a mild increase in mass loading toward lower stellar masses, as well as a strong correlation between mass loading and star-formation-rate surface density, suggesting that more compact starbursts drive more powerful outflows. These trends are consistent with those seen in star-forming galaxies at higher redshifts. Our results highlight the importance of local UV-compact starbursts for understanding feedback processes in low-mass, rapidly star-forming galaxies.

Figures (22)

• Figure 1: (Top panel) Two-dimensional spectrum of SDSS143417 around the H$\alpha$–[ N ii] region, with gray contours tracing flux levels and blue crosses marking the positions used for Gaussian fitting. (Bottom panel) One-dimensional spectra extracted with different apertures: blue (ESO pipeline, 30 px/side, including both knots), orange (IRAF/apall, 5 px centered on the brightest knot), and green (IRAF/apall, asymmetric aperture sampling the secondary knot). Vertical dashed lines indicate the wavelength ranges where the two-dimensional contours reveal asymmetries, serving as guides to identify additional kinematic components.
• Figure 2: Multi-Gaussian fitting of bright emission lines for the galaxies in our sample. The top panels show two-dimensional spectra, with the y-axis in pixel units. Center panels show Gaussian fits of [O iii]$\lambda$5007$\AA$ (Left) and H$\alpha$ (Right). Bottom panels show fit residuals. The flux axis is normalized to the peak emission of each line. Spectra are shown in light blue (Data). The blue shadow represents the variance spectrum. The black line models the fit. The dashed lines show the different fitted components. Grey shadows are flagged regions excluded from fits. The yellow line represents the continuum. The zoom-in insets for the faint line wings are included in the upper-right corner of each plot.
• Figure 3: L-$\sigma$ relationship. The blue and red circles correspond to the narrow and broad components of our sample of LBA galaxies, respectively. Grey symbols are Giant HII regions and HII galaxies at $0 < z< 2.33$ from Terlevich2015.
• Figure 4: Classic diagnostic diagrams based on emission line ratios BaldwinVeilleux_1987 [O iii]$\lambda$5007/$H\beta$ vs. [N ii]$\lambda6584/H\alpha$ (Top) and [O iii]$\lambda$5007/ $H\beta$ vs. [S ii]$\lambda \lambda$6716.6731/$H\alpha$ (Bottom). Symbols show emission lines integrated fluxes (gray squares), narrow (blue circles) and broad (red circles) Gaussian components for each galaxy. The gray dots correspond to the SDSS-DR7 MPA-JHU galaxy sample from Perez-Montero2021. Regions of different excitation mechanisms are labeled and established by theoretical and empirical demarcation lines of Kewley_2001Kewley_2006 (black dashed), Kauffmann_2003 (gray dashed), and Schawinski_2007 (brown solid).
• Figure 5: Outflow mass loading factor as a function of the stellar mass of galaxies (Left) and stellar mass outflow rate as a function of SFR surface density (Right). Our LBAs are shown in red squares. Green and black circles LBGs at $z\sim$ 1.4-3.8 from weldon2024 and llerena2023ionized, respectively, while purple stars are mean-weighted averages from stacks of star-forming galaxies at z$\sim$ 2.1 of Concas_2022. Gray and blue crosses are low-mass star-forming galaxies at $z\sim$ 3-9 observed with JWST/NIRSpec R=2700 spectroscopy from Carniani2024 and Saldana-Lopez2025, respectively. The green crosses are the results from Xu2025 for galaxies at $z\sim3-9$. Yellow stars are local dwarf galaxies of Marasco_2023. Light blue squares indicate local dwarf galaxies from Mcquin_2019. Red dashed lines are rescaled relations from Ilustris-TNG simulations of Nelson_2019 and black dashed lines are from FIRE-2 simulations of Pandya_2021.