An upper limit of 10$^6$ M$_\odot$ in dust from ALMA observations in 60 Little Red Dots

Caitlin M. Casey, Hollis B. Akins, Steven L. Finkelstein, Maximilien Franco, Seiji Fujimoto, Daizhong Liu, Arianna S. Long, Georgios Magdis, Sinclaire M. Manning, Jed McKinney, Marko Shuntov, Takumi S. Tanaka

First listed 2025-05-24 | Last updated 2025-05-24

Abstract

By virtue of their red color, the dust in little red dots (LRDs) has been thought to be of appreciable influence, whether that dust is distributed in a torus around a compact active galactic nucleus (AGN) or diffuse in the interstellar medium (ISM) of nascent galaxies. In Casey et al. (2024) we predicted that, based on the compact sizes of LRDs (unresolved in JWST NIRCam imaging), detection of an appreciable dust mass would be unlikely. Here we present follow-up ALMA 1.3mm continuum observations of a sample of 60 LRDs drawn from Akins et al. (2024). None of the 60 LRDs are detected in imaging that reaches an average depth of $σ_{rms}=22\,μJy$. A stack of the 60 LRDs also results in a non-detection, with an inverse-variance weighted flux density measurement of $S_{1.3mm}=2.1\pm2.9\,μJy$. This observed limit translates to a 3$σ$ upper limit of 10$^6$ M$_\odot$ in LRDs' dust mass, and $\lesssim10^{11}$ L$_\odot$ in total dust luminosity; both are a factor of 10$\times$ deeper than previous submm stack limits for LRDs. These results are consistent with either the interpretation that LRDs are reddened due to compact but modest dust reservoirs (with $A_{V}\sim2-4$) or, alternatively, that instead of being reddened by dust, they have extreme Balmer breaks generated by dense gas ($>10^{9}\,cm^{-3}$) enshrouding a central black hole.

Short digest

ALMA Band-6 (1.3 mm) continuum follow-up of 60 of the reddest and brightest COSMOS-Web little red dots—preselected with MIRI F770W detections—finds no individual detections and a non-detection in the stack. The data reach a mean σ_rms≈22 μJy, with a stacked flux of S_1.3mm=2.1±2.9 μJy, implying 3σ limits of Mdust<10^6 M⊙ and Ldust≲10^11 L⊙, about 10× deeper than previous submm stacks. These results disfavor large, cold dust reservoirs in typical LRDs. They are consistent with compact, modest dust columns (A_V~2–4) or extreme Balmer breaks from very dense gas (n>10^9 cm^-3) enshrouding a central black hole.

Key figures to inspect

• Figure 1: Inspect the JWST cutouts with ALMA contours to verify the absence of >3σ peaks at LRD centroids and to gauge the uniform 1σ RMS (21–23 μJy); also check for any offset positive contours that might hint at extended hosts or unrelated submm sources.
• Figure 2: Read off how the 3σ per-source limit (≈66 μJy) and the stacked 3σ limit (≈10.8 μJy) map onto Mdust and LIR across dust temperatures and redshifts; note where the curves intersect Mdust≈10^6 M⊙ and how they compare to the Setton et al. (2025) constraints.
• Figure 3: Compare the new stacked limit against previous LRD submm/NOEMA results (Labbé 2022; Akins 2024; Xiao 2025; Setton 2025), noting the rescaling to typical LRD luminosities and the consistency with the predicted low-luminosity dust SED; use the overplotted 30 K and 1500 K blackbodies to visualize why large cold…