On the multiplicity of red-Herschel sources and its implications for extreme star formation

TL;DR

This work uses ALMA Band 6 data to quantify the multiplicity of red-Herschel DSFGs in the HARPAS sample, finding a multiplicity fraction of about 20% (potentially ~25% if including close blends) with 2, 3, or 4 components within ~16.6 arcsec. The brightest component typically dominates the flux (64% in doubles, 48% in triples, 42% in quadruples) and full physical association across all components is rare, though many fields contain at least one likely physically associated pair ($Δz<0.01$) in subsets. Deblending with XID+ and photometric-redshift modeling (MMPz) yield median $z_{phot}\approx3.0$, $L_{IR}\approx6.0\times10^{12} L_\odot$, and SFR ~$8.9\times10^{2} M_\odot\,\mathrm{yr}^{-1}$, with substantial redshift uncertainties limiting definitive association claims. Mock DSFG catalogs suggest that while full binding is uncommon, a majority of triples/quadruples harbor at least two physically linked sources, implying these systems can trace overdensities or proto-clusters; the results refine the interpretation of SMG extreme luminosities and motivate follow-up spectroscopy to confirm associations.

Abstract

We study the multiplicity of galaxies in the largest sample of red-Herschel sources ($S_{250 μ\mathrm{m}} < S_{350 μ\mathrm{m}} < S_{500 μ\mathrm{m}}$) using archival ALMA observations. Out of 2416 fields with ALMA detections (from a total of 3,089 analyzed maps), we identify 474 multiple systems within a radius of 16 arcsec (equivalent to the 500 $μ$m Herschel beam-size): 420 doubles, 51 triples, and 3 quadruples. In each case the brightest source contributes, on average, 64, 48, and 42 per cent of the total flux in double, triple, and quadruple systems. The average combined ALMA flux density of the sources in double systems is comparable to that of the two brightest components within triple and quadruple systems. Non-parametric tests suggest that only a small fraction of the double systems ($\lesssim13$ per cent) are comprised of sources with compatible redshifts, while 47-67 per cent of triple and quadruple fields contain at least one potentially associated pair. Simulations using a mock catalogue of dusty star-forming galaxies suggest that 32 per cent of the double systems are likely physically associated ($Δz < 0.01$, i.e. $\lesssim$10 comoving Mpc at $z = 3$) and, while only 8 per cent of the triple and none of the quadruple systems meet this criterion, $\sim$ 70 per cent of them include at least one likely associated pair. Our results suggest that enhanced star formation rates in submillimetre galaxies are primarily driven by internal processes rather than large-scale interactions. This study also provides a catalog of potential overdensities for follow-up observations, offering insights into proto-cluster formation and evolution.

Figures (13)

• Figure 1: ALMA 1.3 mm maps showing examples of the three sub-samples in M-Fields (based on the number of sources above 5$\sigma$ within a detection radius of 16.6 arcseconds): doubles, triples, and quadruples. The values in parentheses indicate the number of fields in each classification out of the 474 M - Fields. The diameter of the ALMA map is $\sim18.6$ arcseconds
• Figure 2: Percentages of the different subsamples of the $\sim$3,000 red-Herschel sources classified from the 1.3 mm ALMA maps. The HARPAS catalogue contains three main categories: S-Fields, PLUM-Fields, and M-Fields. The M-Field classification is further subdivided into doubles, triples, and quadruples.
• Figure 3: Flux density of each component in multiple systems compared to the S-Fields Quiros-Rojas2024. Components are denoted by D$_i$ for doubles, T$_i$ for triples, and Q$_i$ for quadruples. The subindex, $i$, indicates the relative brightness of each component, with 1 for the brightest source. Stars denote the median flux densities, while the error bars represent the 1st and 3rd quartiles of the flux density distribution. The dotted black line indicates the 5$\sigma$ median 1.3 mm detection threshold for the full HARPAS sample, while the dashed red line represents the flux limit proposed by Quiros-Rojas2024, above which 100 per cent of the SMGs are gravitationally amplified. As can be seen, the brighter components ($S_{\mathrm{D1,T1,Q1}}$) in each system (e.g., doubles, triples, and quadruples) exhibit a median flux density comparable to the S-Fields.
• Figure 4: Flux density of the brighter component in each double system ($x$-axis) and its contribution to the total flux density of the system ($y$-axis). The numbers inside the rectangles represent the number of doubles within each flux density bin. The color bar indicates the percentage of doubles in each bin. On average, the brighter component contributes approximately 64 per cent of the total flux across most doubles in each field.
• Figure 5: Cumulative number of galaxies (bottom panels) for the HARPAS sample as a function of Herschel/SPIRE 500 $\mu$m (left) and 1.3 mm (right) flux density and split according to their classification. For the 1.3 mm panel, we include all the individual components within each system, while in the 500 $\mu$m panel we use the original flux density of each field reported in the H-ATLAS catalogue. The top panels, show the fractional contribution to the total of red-Herschel fields observed with ALMA (left) and to the total of ALMA detections (right). Fields with no detections in ALMA tend to be fainter at 500 $\mu$m. We also indicate the threshold proposed in Quiros-Rojas2024 ($S_{1.3\mathrm{mm}} \geq 13.0$ mJy), for identifying gravitationally lensed candidates. This limit remains applicable to multiple systems, as most galaxies in these systems have flux densities below this threshold, with the exception of one, which may exhibit signs of gravitational amplification below our angular resolution ($\sim1$ arcsecond).