Dissecting the Obscured Core of GN20: an Active Galactic Nucleus Outshone by Young Stars

M. Hamed, L. Colina, P. G. Pérez-González, J. Álvarez Márquez, A. Crespo Gómez, L. A. Boogaard, A. Bik, H. Übler, S. E. I. Bosman, A. Alonso-Herrero, M. Perna, S. Arribas, L. Ulivi, M. Annunziatella, L. Costantin, A. Labiano, C. -L. Liao, C. Prieto-Jiménez, P. van der Werf, F. Walter

First listed 2026-07-01 | Last updated 2026-06-30

Abstract

We investigate the relative contributions of star formation and AGN activity to the total energy budget of GN20, one of the most luminous dusty star-forming galaxies known at $z>4$, through spatially resolved spectral energy distribution decomposition. We perform Bayesian SED fitting with CIGALE on two spatially distinct apertures: the nuclear core (r=0.14", $\sim$1kpc physical) and the full galaxy (r=1.4", 9.9 kpc), combining JWST/NIRCam and MIRI broadband imaging, JWST/NIRSpec PRISM IFU pseudo-continuum photometry spanning 42 wavelength bins across rest-frame $0.12$--$1.05μ$m, and archival HST and millimeter interferometry data from NOEMA and PdBI. The integrated SED is dominated by stellar-heated dust, with only a marginal AGN contribution at galaxy-wide scales ($f_\mathrm{AGN}^\mathrm{int}=0.09\pm0.02$). The nuclear core, however, requires a significant AGN component ($f_\mathrm{AGN}=0.34\pm0.05$) to account for a mid-infrared excess at rest-frame $\sim$2.5--3.6$μ$m characteristic of AGN-heated torus dust. The AGN accounts for $\sim34\%$ of the nuclear infrared luminosity but only $\sim9\%$ of the total integrated $L_\mathrm{IR}$, explaining its weak signature in integrated diagnostics and its consistency with existing upper limits from Spitzer spectroscopy. The inferred black hole mass places GN20 within the local $M_\mathrm{BH}$--$M_\mathrm{bulge}$ relation at the Eddington limit, and in the overmassive regime at sub-Eddington accretion rates, suggesting early and rapid black hole assembly concurrent with the dominant starburst. GN20 exemplifies a class of systems where nuclear-scale SED decomposition, enabled by the angular resolution and infrared sensitivity of JWST, is the only means to uncover a buried AGN overwhelmed by galaxy-wide star formation.

Short digest

This paper uses spatially resolved CIGALE SED fitting to separate the ∼1 kpc nuclear core of GN20 from the full 9.9 kpc galaxy, combining JWST NIRCam, MIRI, NIRSpec PRISM pseudo-continuum, HST, and millimeter data. The central result is that the galaxy-wide SED is overwhelmingly star-formation powered, with only a marginal integrated AGN fraction of 0.09 ± 0.02, while the nucleus requires a much stronger AGN contribution of 0.34 ± 0.05 to explain a rest-frame 2.5–3.6 μm mid-infrared excess from torus-heated dust. That means the AGN contributes about one-third of the nuclear infrared luminosity but only about 9% of the total L_IR, naturally explaining why integrated diagnostics and earlier Spitzer limits made the source look starburst-dominated. The broader significance is that GN20 looks like a buried-black-hole case where only JWST-resolution, nuclear-scale decomposition can uncover ongoing early SMBH growth inside an extreme z = 4.055 dusty starburst.

Key figures to inspect

• Figure 1. Use this to show the measurement geometry that drives the paper’s main result: the compact nuclear aperture at r = 0.14 arcsec versus the much larger integrated aperture at r = 1.4 arcsec. The wavelength-dependent channel maps also make clear that the authors are isolating a distinct core within a much more extended system, which is essential context for why integrated photometry misses the AGN.
• Figure 2. This figure is a clean observational demonstration that the core and integrated SED shapes differ before any modeling is imposed. It highlights the core non-detections at shorter wavelengths and the relative rise toward the mid-infrared, which motivates the claim that a buried nuclear component is emerging only at longer rest-frame wavelengths.
• Figure 3. This is the paper’s key evidence figure because it directly compares the best-fit decomposed SEDs for the nucleus and the whole galaxy and shows that the core needs an AGN component while the integrated aperture does not. The no-AGN residuals in the core, especially around the F1280W and F1800W excess, are the most compact visualization of the conclusion that torus-heated dust is required in the center of GN20.
• Figure 4. Include this later comparison figure because it places GN20’s core mid-infrared excess in the context of other extreme high-redshift dusty or AGN-hosting systems. It matters for the digest because it connects the GN20 nuclear SED to a broader population of obscured, rapidly growing black holes whose AGN signatures would be diluted or missed in unresolved galaxy-scale measurements.