Clump-Scale Dust Attenuation in Epoch of Reionization Galaxies: Spatially Resolved Properties from FirstLight Simulations

TL;DR

This study investigates dust attenuation and re-emission in clumpy galaxies during the epoch of reionization (z ≈ 6–9) using cosmological zoom-in simulations from the FirstLight suite coupled with post-processing radiative transfer via SKIRT. It reveals pronounced spatial variation: star-forming clumps host higher dust column densities and well-mixed dust-star geometries, yielding grayer attenuation curves and larger IRX compared to the diffuse and system-integrated components. A novel IRX-$\Delta\beta$ toy model with two parameters, the dust-to-star scale-height ratio $R$ and the fiducial UV optical depth $\tau^{\rm fid}_{\rm UV}$, provides a quantitative handle on disentangling optical depth from geometry, reproducing the clump versus system-integrated trends and aligning with REBELS-IFU observations for $z\sim7$ systems. The findings imply that pixel-by-pixel attenuation laws and spatially resolved IRX-$\Delta\beta$ maps are crucial for accurate SED fitting and for interpreting the physics of dust in the early universe, with broad relevance for JWST/ALMA observations and future high-resolution studies.

Abstract

Understanding dust attenuation in galaxies at both integrated and spatially resolved scales is fundamental for accurately determining the physical properties of galaxies. Recent high-spatial-resolution observations with ALMA and JWST enable investigations of spatially resolved properties in high-redshift galaxies ($z \gtrsim 6$), but spatial variations in dust properties remain poorly constrained. We use cosmological zoom-in simulations combined with post-processing dust radiative transfer calculations for 376 clumpy galaxies at $z=6$-$9$ with stellar masses of $M_* \gtrsim 10^9 \, M_\odot$. For each system, we investigate dust attenuation and re-emission properties for three components: system-integrated, individual clumps, and diffuse regions. We find that system-integrated attenuation curves are grayer than the Calzetti curve, even when assuming MW- or SMC-type dust. Attenuation curves of individual clumps are even grayer, while diffuse regions exhibit steeper curves owing to enhanced scattering in optically thin environments. Since the effects of optical depth and dust-star geometry are intrinsically degenerate in attenuation curves, we introduce a toy model based on the IRX-$Δβ$ plane, where $Δβ$ denotes the difference between attenuated and intrinsic UV slopes. Applying this framework, we find that clumps have dust column densities approximately an order of magnitude higher than system-integrated values and exhibit co-spatial or dust-extended geometries. In contrast, system-integrated attenuation reflects star-extended geometries driven by contributions from optically thin diffuse regions. We apply this framework to REBELS-IFU galaxies at $z \sim 7$ and find good agreement with our simulation predictions.

Figures (12)

• Figure 1: Histogram of clump properties. The top panel shows the mass-weighted stellar age distributions for each component: clumps (cyan), diffuse (orange), and system-integrated (black). The bottom panel shows the surface density distributions of SFR (cyan) and stellar mass (blue) for clump components. Note that these surface densities are only well-defined for clumps due to their compact nature.
• Figure 1: SEDs of FL957 at $z=7.7$ computed with different SKIRT settings. In the upper panel, we plot the SED for the fiducial case and two runs with more wavelength bins ($n_\lambda = 300$), and more photon packages ($n_{\rm p} = 10^8$) launched per wavelength bin. Additionally, we show the outcome of simulations for which stochastic heating or self-absorption is turned off ("LTE’’, "w/o self-absorption’’). The detailed SKIRT settings in each model are described in Table \ref{['table:convergence_test']}. In the lower panel, we show the fractional difference between the various runs and the fiducial model.
• Figure 1: Comparison of dust properties between system-integrated values and individual components for all identified clumpy galaxies at $z=6-9$ as the same as Figure \ref{['fig:dust_properties_statistical']}.
• Figure 2: Projected distribution of simulated galaxy FL957 at $z=7.7$, one of the identified clumpy galaxies. Upper panels: (a) projected gas number density, (b) surface SFR density with yellow lines identifying star-forming clumps (see Section \ref{['subsec:clump_identification']}), (c) dust column density, and (d) mass-weighted dust temperature. Lower panels: (e) rest-frame UV surface brightness ($\nu L_\nu$ at 1600 Å), (f) total IR emission integrated over $\lambda_{\rm rest} = 8$-$1000\,\mu{\rm m}$, (g) mock JWST/NIRCam three-color image (F115W, F200W, F356W) without dust attenuation, and (h) the same with dust attenuation. Each panel shows a 10 kpc$\times$ 10 kpc region with a projected depth of 10 kpc.
• Figure 2: Dust attenuation curves (top panel) and attenuation curve slope ($S \equiv A_{\rm UV}/A_{\rm V}$) versus V-band dust attenuation ($A_{\rm V}$) for all identified clumpy galaxies at $z=6-9$ (bottom panel) as the same as Figure \ref{['fig:attenuation_curve']}. The SMC extinction curve is adopted from Gordon:2003.