JWST finds where a Little Red Dot’s red rise stops

Little Red Dots earn their name from their compact appearance and unusually red optical light. Yet they also have a faint blue ultraviolet component, and many show broad hydrogen emission lines associated with rapidly moving gas. Different combinations of stars, dust and black-hole activity can reproduce parts of this pattern. The challenge is finding an explanation that satisfies all the evidence together.

In a preprint submitted to MNRAS, E. Iani and colleagues examine Abell2744-QSO1, a Little Red Dot at redshift 7.04. They find that its steep increase in brightness toward redder optical wavelengths stops around the beginning of the near-infrared. That change exposes weaknesses in several proposed models, rather than delivering a single winner.

Following the light farther into the infrared

The foreground galaxy cluster Abell 2744 gravitationally lenses QSO1, magnifying it into three images of the same source. The researchers observed all three with JWST’s Mid-Infrared Instrument, MIRI, using filters centred on observed wavelengths of 7.7 and 10 micrometres.

Cosmic expansion stretches the light during its journey. At QSO1’s redshift, those measurements probe wavelengths of about 0.96 and 1.24 micrometres as emitted by the source: its rest-frame near-infrared. This extends the previously observed optical continuum, the underlying light between emission lines.

The team combined these measurements with existing Hubble and JWST imaging, JWST spectroscopy and longer-wavelength non-detections. They subtracted nearby contaminating light, corrected for lensing magnification and allowed for residual differences between the three images. QSO1 remained unresolved in MIRI; the most restrictive conservative limit puts its near-infrared half-light radius below 64 parsecs.

A turnover, not an endless red rise

Across the two MIRI bands, QSO1’s flux density per unit frequency is approximately flat, unlike its steep optical rise. This is a change across wavelengths, not evidence that the object has faded over time. Estimated emission-line contributions of about 3% and 7% in the two filters are too small to explain the flattening.

An idealised thermal spectrum—a blackbody—at about 4,450 kelvin describes the optical and near-infrared continuum. But this does not establish a unique physical temperature or emitting surface. A hotter blackbody dimmed and reddened by dust fits comparably well.

There is also no detected upturn from the hot dust commonly seen in unobscured quasars within the measured range. The important boundary is 1.24 micrometres in the source’s rest frame: these observations do not show what happens immediately beyond it.

Models that fit the colours can fail the mass test

The researchers compared stellar populations, dust-reddened black-hole emission and accretion discs surrounded by dense gas. Their statistically preferred joint fits to the spectrum and imaging produce the red continuum with roughly three to four billion solar masses of stars.

Earlier measurements of gas motion, however, indicate a black hole of about 50 million solar masses and leave room for at most about 20 million solar masses in host stars. The preferred stellar solutions exceed that host limit by roughly 140–200 times. If the dynamical constraint holds, those solutions cannot describe the source.

Restricting the stellar mass changes the fit: an accretion disc and dense gas cocoon then supply most of the red light. But this model predicts too much light in the longer-wavelength MIRI band, a mismatch of about five times the quoted measurement uncertainty.

A different family of models lets dense gas absorb and re-emit the accretion energy as a nearly thermal spectrum, forming an apparent emitting surface called a pseudo-photosphere. These models reproduce the turnover, but match the luminosity only with a black hole roughly 300 times less massive than the dynamical estimate.

These failures do not rule out every dense-gas explanation. The cocoon calculations simplify gas structure and radiation transport, while the thermal-envelope models assume particular accretion conditions. The authors also caution that statistical model rankings are affected by modelling limitations and correlations between the spectrum and imaging.

Dust could still hide the energy

Dust reddening remains possible, but absorbed energy must emerge elsewhere. Under the reddened-disc model and a full-covering dust assumption, the team’s one-standard-deviation limit allows dust at 1,200 kelvin to re-emit only about 2% of the inferred absorbed luminosity. Most could instead escape detection through cooler dust at roughly 70–600 kelvin.

That is an allowed energy budget, not a detection of warm dust. It depends on the assumed spectrum, dust properties and geometry; partial covering relaxes the requirements.

Taken together, the dynamics favour a source powered by accretion onto a black hole, but the new observations cannot distinguish dust reddening from processing by dense gas. Deeper MIRI observations at observed wavelengths around 18 micrometres and beyond could reveal whether the continuum rises again and tighten the dust budget. For now, the turnover is measured more securely than the mechanism behind it.