A Redshift-based Red Selection of Dusty Star-forming Galaxies

A. J. Barger, L. L. Cowie, S. J. McKay, F. E. Bauer

First listed 2026-05-26 | Last updated 2026-05-22

Abstract

We use JWST observations (1.5 micron to 4.44 micron), together with complete ALMA observations (870 micron and/or 1.2 mm), of the massive lensing cluster field A2744 to show that galaxies between z=1.5 and z=5.5 with rest-frame red colors f_J/f_V > 3 correspond to dusty star-forming galaxies (DSFGs), little red dots (LRDs), and quiescent galaxies. The color selection picks out 34 of the 41 >4.5-sigma ALMA sources in the field (83%). We find that the luminous red sources are generally extended, while the less luminous red sources are almost all compact and correspond to the LRD population. We also find that the great majority of the luminous, extended red sources are DSFGs based on the ALMA data, with a small admixture of quiescent galaxies at z<3-4 that we identify based on their location in the rest-frame U-V versus V-J diagram. We do not detect any LRDs or quiescent galaxies at the >3-sigma level in the ALMA images. Roughly 10% of the DSFGs have high rest-frame X-ray luminosities and must be AGN dominated. The DSFGs and quiescent galaxies nearly all have M_star>10^{10} solar masses. These massive galaxies become rare at z>5, paralleling the fall off in the number of detected DSFGs.

Short digest

Using JWST 1.5–4.44 μm photometry together with complete 870 μm and 1.2 mm ALMA coverage in A2744, this paper shows that a simple rest-frame red cut, f_J/f_V > 3, efficiently selects dusty systems at z=1.5–5.5, recovering 34 of the 41 >4.5σ ALMA sources in the field. Within that red-selected sample, the bright sources are mostly extended dusty star-forming galaxies, while the fainter compact members are overwhelmingly little red dots. A smaller lower-redshift component enters the sample as quiescent galaxies, which the authors separate in UVJ space, and neither the LRDs nor the quiescent systems are detected significantly by ALMA. The DSFG and quiescent populations are largely massive, with M_star > 10^10 solar masses, and both they and the ALMA-detected DSFGs become scarce beyond z > 5, while only about 10% of the DSFGs appear clearly AGN dominated from their X-ray luminosities.

Key figures to inspect

• Figure 5. This is the key selection figure because it defines the rest-frame red cut against 1.22 μm luminosity across redshift and shows where the ALMA detections, LRDs, compact sources, and X-ray-bright objects fall relative to the adopted thresholds. It is the cleanest single view of why the authors argue that f_J/f_V > 3 is an efficient DSFG selector while still containing distinct compact LRD and AGN subsets.
• Figure 6. This figure directly supports the paper’s main physical split by showing source size versus NIR luminosity for the red-selected population. It makes clear that luminous red galaxies are generally extended, consistent with the ALMA-detected DSFG population, whereas compact red sources cluster at lower luminosities and align with the published LRDs.
• Figure 8. Use this UVJ diagram to understand the paper’s main caveat and classification step: not every red source is dusty. The figure shows that a small subset of red galaxies, especially at lower redshift, falls in the quiescent box, while most ALMA-detected red sources remain outside it and are therefore interpreted as dusty star-forming systems.
• Figure 9. This is the best late-stage synthesis figure because it tracks the relative numbers of DSFGs, LRDs, and candidate quiescent galaxies within the red-selected sample as a function of redshift. It captures the paper’s broader conclusion that massive DSFGs and quiescent galaxies fade rapidly above z≈5 while LRDs become comparatively more prominent.
• Figure 4. This X-ray versus redshift diagnostic grounds the statement that only a minority of the DSFG sample is AGN dominated. By marking which ALMA sources exceed the adopted high rest-frame 8–28 keV luminosity thresholds, it isolates the small X-ray-bright tail that requires substantial AGN power.