Early galaxy growth: mergers or gravitational instability?

We investigate the spatially resolved morphology of galaxies in the early Universe. We consider a typical redshift z = 6 Lyman break galaxy, `Althaea', from the SERRA hydrodynamical simulations. We create mock rest-frame ultraviolet (UV), optical, and far-infrared observations, and perform a two-dimensional morphological analysis to deblend the galaxy disc from substructures (merging satellites or star-forming regions). We find that the [C II]158 mu m emitting region has an effective radius 1.5-2.5 times larger than the optical one, consistent with recent observations. This [C II] halo in our simulated galaxy arises as the joint effect of stellar outflows and carbon photoionization by the galaxy UV field, rather than from the emission of unresolved nearby satellites. At the typical angular resolution of current observations (greater than or similar to 0.15 arcsec) only merging satellites can be detected; detection of star-forming regions requires resolutions of less than or similar to 0.05 arcsec. The [CII]-detected satellite has a 2.5-kpc projected distance from the galaxy disc, whereas the star-forming regions are embedded in the disc itself (distance less than or similar to 1 kpc). This suggests that multicomponent systems reported in the literature, which have separations greater than or similar to 2 kpc, are merging satellites, rather than galactic substructures. Finally, the star-forming regions found in our mock maps follow the local L-[C II]-SFRUV relation of galaxy discs, although sampling the low-luminosity, low-SFR tail of the distribution. We show that future James Webb Space Telescope observations, bridging UV and [C II] data sets, will be exceptionally suited to characterize galaxy substructures, thanks to their exquisite spatial resolution and sensitivity to both low-metallicity and dust-obscured regions that are bright at infrared wavelengths.