The James Webb Space Telescope has revealed a population of compact, reddish galaxies in the early universe: little red dots (LRDs), whose emission has a characteristic V-shape: a red optical continuum and a blue ultraviolet excess. A similar spectral shape is seen in rare blue hot dust-obscured galaxies (BHDs), a subclass of hyperluminous obscured quasars at z~1–4. In those, the blue excess is explained by scattering of active galactic nucleus light off dust escaping a dense cocoon. A natural question arises: are LRDs highly reddened relatives of BHDs, that is, are we seeing the same evolutionary phase at different cosmic epochs? This work settles the debate by comparing spectral energy distributions, morphology, X-ray properties, and number densities of both populations.
The authors assembled a sample of 326 LRDs with V-shaped spectra, selected from JWST/NIRCam data in the COSMOS-Web and PRIMER surveys, as well as from the catalog of Kokorev et al. (2024). For comparison, they used 26 BHDs from the 'nuclear' sample of hot dust-obscured galaxies with confirmed redshifts. Both datasets were fitted with a linear combination of active galactic nucleus and host galaxy templates from the Assef et al. (2010) library, varying dust attenuation (AV) for two components: a heavily obscured one (optical/near-IR) and a nearly transparent one (UV excess). Additionally, they compared median bolometric luminosities, effective radii, X-ray detection rates from Chandra and XMM-Newton, and estimated number densities via luminosity functions.
The V-shape in LRDs and BHDs arises at different wavelengths: for BHDs the 'bottom' of the valley falls at longer wavelengths due to strong dust attenuation (AV~20–60 mag), whereas for LRDs the attenuation is moderate (AV~2.6±1.4 mag). The blue excess in LRDs cannot be scattered light from the active nucleus as in BHDs: with such weak obscuration, the scattered light fraction would be negligible, yet observations show that in LRDs it is comparable to BHDs (~1%). Moreover, LRDs lack the hot dust characteristic of BHDs: in the mid-infrared (MIRI F1800W) LRDs are barely detected (only 6% versus expected brightness for BHDs), and their F770W–F1800W color is bluer by ~3 mag. X-ray observations also differ: among LRDs only two objects out of hundreds are detected in soft X-rays with moderate absorption (NH~$10^{22}$–$10^{23}$ cm^-2), whereas among BHDs five out of 44 observed are detected with column densities above $10^{23}$ cm^-2. Finally, LRDs are compact (effective radius ~100 pc), while BHDs are extended (1–4 kpc). All this indicates that LRDs are a distinct population, not a highly reddened analog of BHDs.
This work closes the hypothesis of a direct evolutionary link between LRDs and BHDs. If in BHDs the blue excess is light from the active nucleus scattered by a dusty outflow, then in LRDs it likely originates from stars in a compact host galaxy, not from an accreting black hole. This means that the growth mechanisms of supermassive black holes in the early universe differ from those at later epochs: LRDs likely represent a phase of rapid accretion onto black hole seeds in gas-rich but dust-poor systems, whereas BHDs are a late stage of massive galaxy mergers with a powerful dusty cocoon. This distinction is important for understanding how the first black holes gained mass within the first billion years.
Further observations with JWST/MIRI in the mid-infrared and deep X-ray surveys (e.g., with the future AXIS mission) will allow more precise measurements of the fraction of LRDs with hot dust and X-ray emission. High-resolution spectroscopy will help separate the contribution of stars and the active nucleus to the UV excess. Systematic searches for highly reddened BHD analogs (like the object Virgil at z=6.6) using mid-infrared color selection criteria will also be needed to understand how rare such systems are in the early universe and their connection to LRDs.
The results will impact models of supermassive black hole formation and the interpretation of JWST demographic data on early galaxies.
Conduct deep spectroscopy of the LRD sample in the mid-infrared and X-rays, and expand the search for high-redshift BHD-like objects.
This work directly addresses the unsolved problem of the origin of supermassive black holes: how seeds of 10^2–10^5 solar masses grew to 10^9 solar masses within the first 800 million years. The differing accretion pathways in LRDs and BHDs indicate multiple growth mechanisms, consistent with models of super-Eddington accretion and massive primordial seeds, but requires refining the role of mergers and feedback.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The authors compared samples of LRD (326 objects) and BHD (26 objects), fitting their spectral energy distributions with a linear combination of active galactic nucleus and host galaxy templates with separate extinction. They found that the blue excess in LRD cannot be explained by scattered light from the active nucleus, as in BHD, due to moderate extinction (AV~2.6 vs. 20–60), the absence of hot dust in the mid-infrared, and the low rate of X-ray detections. This points to different mechanisms producing the V-shaped spectra and different growth pathways for supermassive black holes in the early Universe.
Where its strength is
The work clearly separates two superficially similar classes of objects using quantitative criteria (extinction, hot dust, X-rays) and demonstrates that LRD are more likely compact galaxies with moderate star formation rather than obscured quasars. This provides a foundation for further study of the nature of LRD.
Where it could go next
Worth trying
Why this may matter
Understanding the nature of LRD is important for building a coherent picture of supermassive black hole growth and galaxy evolution in the first billion years after the Big Bang.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.57, against a usual 0.64. It looks like few have drilled nearby.
Among LRDs, only about 6% are detected in the mid-infrared at 18 microns, whereas in BHDs this region dominates the bolometric luminosity: a difference of 3 magnitudes in color.
In Carl Sagan's novel 'Contact', the radio signal comes from a source hidden by dust clouds, echoing the idea that supermassive black holes in the early universe may be completely hidden from us by dust until they begin to scatter light.