Using ALMA to resolve the nature of the early star-forming large-scale structure PLCK G073.457.5
Key Words.:Large-scale structure of Universe – Submillimetre: galaxies – Radio continuum: galaxies – Radio lines: galaxies – Galaxies: star formation
Galaxy clusters at high redshifts are key targets for understanding matter assembly in the early Universe, yet they are challenging to locate. A sample of more than 2000 high- candidate structures have been found using Planck’s all-sky submillimetre maps, and a subset of 234 have been followed up with Herschel-SPIRE, which showed that the emission can be attributed to large overdensities of dusty star-forming galaxies. However, the individual galaxies giving rise to the emission seen by Planck and Herschel have not yet been resolved nor characterized, so we do not know whether they constitute the progenitors of present-day, massive galaxy clusters. In an attempt to address this, we targeted the eight brightest Herschel-SPIRE sources in the centre of the Planck peak G073.457.5 using ALMA at 1.3 mm, and complemented these observations with multi-wavelength data from Spitzer-IRAC, CFHT-WIRCam , and JCMT’s SCUBA-2 instrument. We detect a total of 18 millimetre galaxies brighter than 0.3 mJy in 2.4 arcmin. The ALMA source density is 8–30 times higher than average background estimates, and larger than seen in typical “proto-cluster” fields. We are able to match all but one ALMA sources to their NIR counterparts. The most significant (four) SCUBA-2 sources are not included in the ALMA pointings, but we find an 8 stacking detection of the ALMA sources in the SCUBA-2 map at 850 m. We derive photometric redshifts, IR luminosities, star-formation rates, stellar masses, dust temperatures, and dust masses for all the ALMA galaxies; the photometric redshifts are concentrated around and the near-IR colours show a “red” sequence, while the star-formation rates indicate that one third of the galaxies are “starbursts,” the others being main-sequence star-forming galaxies. Serendipitous CO line detections of two of the galaxies appear to match their photometric redshifts, with . We thus find that the Planck peak G073.457.5 contains a proto-cluster candidate at , and one of the richest regions of ALMA sources known. The ALMA-detected members are massive galaxies on the main-sequence relation and account for a total star-formation rate of at least .
Hierarchical clustering models of large-scale structure and galaxy formation predict that the progenitors of the most massive galaxies in today’s clusters are dusty star-forming galaxies at high redshift ( 2-3, e.g., Lilly+99; Swinbank+08). Observationally, this picture is supported by the clustering measurements (Blain+04) of submillimetre galaxies (SMGs), and by their relative abundance and distribution in known proto-clusters (e.g., Capak+11; Hayashi+12; Casey+15; Hatch+16; Overzier2016), while recent studies may also indicate a complicated evolution (e.g., Hayashi+2017). High-redshift structure formation studies in millimetre (mm) and submillimetre (submm) wavelength ranges have the advantage of providing access to high redshifts by utilizing the steep rise in the warm dust spectrum of infrared galaxies (the “negative k-correction,” BlainLongair93 BlainLongair93; see also Guiderdoni+97 Guiderdoni+97) and can build on an observed correlation between the total matter density and the cosmic infrared background fluctuations (PlanckXVIII14).
Substantial progress has been made on probing the early formation of massive structures and galaxy clusters through mm/submm observations (see Casey16, for a recent discussion), with a strong emphasis on main-sequence evolution versus starbursts and mergers (see also Narayanan+15). Mechanisms for rapid, episodic bursts, suggested to explain how the member galaxies are assembled and grow during cluster formation, can be tested with measurements of mm galaxy number densities and gas depletion timescales in cluster-forming environments. Likewise, the processes responsible for triggering star formation that is coherent over large spatial scales may depend on environmental effects, which can only be tested using a variety of high quality data over wide areas.
The Planck satellite mapped out the whole sky between 30 and 857 GHz with a beam going down to (PlanckI14), giving it the capability of detecting the brightest mm/submm regions of the extragalactic sky at Mpc scales. A component-separation procedure using a combination of Planck and IRAS data was applied to the maps outside of the Galactic mask to select over 2000 of the most luminous submm peaks in the cosmic infrared background (CIB), with spectral energy distributions peaking between 353 and 857 GHz (PlanckXXXIX15, the “PHz” catalogue). 234 of these peaks (chosen such that at 545 GHz, , and ) were subsequently followed up with Herschel-SPIRE observations between 250 and 500 m, and the half-arcminute (or better) resolution was capable of distinguishing between bright gravitational lenses and concentrations of clustered mm/submm galaxies around redshifts of 2–3 (PlanckXXVII15). Here, we present the first detailed mm analysis of one of these highly clustered regions, PLCK G073.457.5 (hereafter G073.457.5), which was observed with ALMA in Cycle 2. We combine NIR and FIR multi-wavelength data with the resolving power of ALMA to identify the individual galaxies responsible for much of the Planck submm flux and to constrain their physical properties.
This paper on G073.457.5 is structured as follows. In Sect. 2 we re-capitulate the features of the Planck/Herschel sample, followed by Sect. 3, where we present details of the ALMA observations, data reduction, and results. In Sect. 4 we describe the set of multi-wavelength data on G073.457.5, comprising Herschel-SPIRE, SCUBA-2, Spitzer-IRAC, and CFHT-WIRCam observations. In Sect. 5 we present the analysis of these data, where we estimate the mm galaxy number density of G073.457.5, derive photometric redshifts and IR properties of each galaxy (such as dust temperature, dust mass, IR luminosity, star-formation rate and stellar mass), and in Sect. 6 we interpret serendipitous line detections. In Sect. 7 we discuss our findings and synthesize the interpretation. The paper is then concluded in Sect. 8.
Throughout this paper we use the parameters of the best-fit Planck flat CDM cosmology (PlanckXIII15), i.e., , and note that in this model 1 at corresponds to a physical scale of 8.7 kpc.
2 The Planck/Herschel high-z sample
A dedicated Herschel (Pilbratt+10) follow-up programme with the SPIRE instrument for 234 Planck targets (PlanckXXVII15) found a significant excess of “red” sources (where “red” means / and / (consistent with 2), in comparison to reference SPIRE fields. Assuming a single common dust temperature for the sources of = 35 K, IR luminosities of typically L were derived for each SPIRE source, yielding star-formation rates (SFRs) of around 700 M yr. If these observed Herschel overdensities are coherent structures, their total IR luminosity would peak at L, or in terms of SFR, at M yr, i.e., the equivalent of ten typical sources constituting to the overdensity.
A small subset of 11 Herschel sources are now known to be gravitationally lensed single galaxies (Canameras+15), including the extremely bright G244.8+54.9, greater than 1 Jy at 350 m. ALMA data for such sources, also aided by HST-based lensing models, have enabled extremely detailed studies of high- star-forming galaxies (e.g., Nesvadba+16; Canameras+17a; Canameras+17b).
In a recent paper, MacKenzie+16 presented SCUBA-2 follow-up of 61 Planck/Herschel targets at 850 m, each observation covering essentially the full emission of the Planck peak. 172 sources were detected in the maps with high confidence (), and by fitting modified blackbody dust SEDs it was shown that the distribution of photometric redshifts peaks between and .
Further studies of the Planck/Herschel targets, based on NIR and optical data, have also been carried out (e.g., G95.561.6 by FloresCacho+16) or are in progress (e.g., a sample by Martinache16) with the aim of characterizing these sources; indeed, FloresCacho+16 was able to conclude that G95.561.6 consists of two significantly clustered regions at and at , which further motivates their utility for studying high-redshift clustering.
In the current paper we focus on directly detecting the galaxies responsible for giving rise to the Planck peak using the high-resolution (sub-)mm imaging capabilities of ALMA. Our target G073.457.5 was included in the Planck/Herschel sample from the selection of the first public release data of the Planck Catalogue of Compact Sources111Note, that for the latest Planck release(PlanckXXXIX15), G073.457.5 lies just inside the more conservatively applied Galactic mask. with a 545-GHz (PSF FLUX) flux density of mJy. It was chosen for an ALMA proposal, together with three other Planck/Herschel targets, based on the high overdensity of Herschel sources within the Planck contour (see Fig. 1) and the availability of additional NIR and submm data.
3 ALMA observation of G073.457.5
We received 0.4 hours of on-source observing time on G073.457.5 with ALMA in Cycle 2 (PID: 2013.1.01173.S, PI R. Kneissl). We targeted the eight sources found in the SPIRE field that were consistent with a “red” colour, within the error bar, as defined above. A standard Band 6 continuum set-up around 233 GHz (1.3 mm) was used, with four 1.78-GHz spectral windows divided into the two receiver sidebands, separated by 16 GHz (i.e., central frequencies of 224, 226, 240 and 242 GHz). 34 antennas were available in the array configuration during the time of the observation, and the resulting synthesized beam achieved an angular resolution of (FWHM) with a position angle of . The central sensitivity was mJy beam in all eight fields (see Fig. 2 for an overview, and note the Herschel SPIRE IDs, as given in Table LABEL:table:2); with this sensitivity ALMA can detect all SPIRE sources at any redshift assuming a dust temperature of K and that all the SPIRE flux comes from a single source, since the detection significance increases at higher redshifts and with higher temperatures. The observatory standard calibration was used with J2232+1143, a grid monitoring source, as bandpass calibrator, Ceres as additional flux calibrator, and all pointings in this data set share the same phase calibrator, J23060459. The single pointings were convolved with the primary antenna beam pattern (roughly Gaussian with , assuming ).
The data were reduced with standard CASA tasks (CASA+2007), including de-convolution, to yield calibrated continuum images with flat noise characteristics for source detection. A mask was applied to the beam-uncorrected maps with a 2 CLEAN threshold, yielding 13 sources in six fields, where the detection is based on the peak pixel surface brightness. In addition, the single brightest sources from each of the remaining two fields were included in the sample, since they were both well centred, with . During cross-matching with Spitzer maps, three additional sources were identified with . The final sample, containing 18 ALMA sources with flux densities 0.3 mJy and , is presented in Table 3.
The flux density results were derived from applying ImageFitter to the CLEANed maps and integrating over each source. They are presented in Table 3, along with the angular sizes for nine sources that were best fit with an extended profile (and four of which had a major axis determined with ). In the nine remaining cases the fit for size did not converge well and these are listed as point sources. In addition, for each source we give the peak flux density at 233 GHz from the beam de-convolved map (which is more accurate for the nine point sources) and the coordinate for the position of the peak surface brightness. Note that ALMA source ID 16, which is on the edge of pointing field 7, has a recovered peak flux density of , i.e., 3.5, and should thus be considered tentative, in spite of the match with a Spitzer source (see next section, Fig. LABEL:Fig2).