Very compact millimeter sizes for composite star-forming/AGN submillimeter galaxies
Soh Ikarashi, Karina Caputi, Kouji Ohta, R. J. Ivison, Claudia D. P. Lagos, Laura Bisigello, Bunyo Hatsukade, Itziar Aretxaga, James S. Dunlop, David H. Hughes, Daisuke Iono, Takuma Izumi, Nobunari Kashikawa, Yusei Koyama, Ryohei Kawabe, Kotaro Kohno, Kentaro Motohara, Kouichiro Nakanishi, Yoichi Tamura, Hideki Umehata, Grant W. Wilson, Kiyoto Yabe, Min S. Yun
aa r X i v : . [ a s t r o - ph . GA ] O c t Draft version October 31, 2017
Typeset using L A TEX twocolumn style in AASTeX61
VERY COMPACT MILLIMETER SIZES FOR COMPOSITE STAR-FORMING/AGN SUBMILLIMETERGALAXIES
Soh Ikarashi, Karina I. Caputi, Kouji Ohta, R. J. Ivison,
Claudia D. P. Lagos, Laura Bisigello,
1, 6
Bunyo Hatsukade, Itziar Aretxaga, James S. Dunlop, David H. Hughes, Daisuke Iono,
9, 10
Takuma Izumi, Nobunari Kashikawa,
Yusei Koyama,
9, 10,12
Ryohei Kawabe,
9, 10
Kotaro Kohno,
7, 13
Kentaro Motohara, Kouichiro Nakanishi,
Yoichi Tamura, Hideki Umehata, Grant W. Wilson, Kiyoto Yabe, andMin S. Yun Kapteyn Astronomical Institute, University of Groningen, P.O. Box 800, 9700 AV Groningen, Netherlands Department of Astronomy, Kyoto University, Kitashirakawa-Oiwake-Cho, Sakyo-ku, Kyoto 606-8502, Japan Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, UK European Southern Observatory, Karl Schwarzschild Str. 2, D-85748 Garching, Germany International Centre for Radio Astronomy Research, University of Western Australia, 7 Fairway, Crawley 6009, Perth WA, Australia SRON Space Research of Netherlands, 9747 AD, Groningen, The Netherlands Institute of Astronomy, University of Tokyo, 2-21-1 Osawa, Mitaka, Tokyo 181-0015, Japan Instituto Nacional de Astrof´ısica, ´Optica y Electr´onica (INAOE), Luis Enrique Erro 1, Sta. Ma. Tonantzintla, 72840 Puebla, Mexico National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan SOKENDAI (The Graduate University for Advanced Studies), Shonan Village, Hayama, Kanagawa 240-0193, Japan Optical and Infrared Astronomy Division, National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan Subaru Telescope, 650 North Aohoku Place, Hilo, HI 96720, USA Research Center for the Early Universe, School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113-0033, Japan Division of Particle and Astrophysical Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602,Japan The Open University of Japan, 2-11 Wakaba, Mihama-ku, Chiba 261-8586, Japan Department of Astronomy, University of Massachusetts, Amherst, MA 01003, USA Kavli Institute for the Physics and Mathematics of the Universe (WPI), The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba277-8583, Japan
ABSTRACTWe report the study of far-IR sizes of submillimeter galaxies (SMGs) in relation to their dust-obscured star formationrate (SFR) and active galactic nuclei (AGN) presence, determined using mid-IR photometry. We determined themillimeter-wave ( λ obs = 1100 µ m) sizes of 69 ALMA-identified SMGs, selected with ≥ σ confidence on ALMAimages ( F µ m = 1 . z = 1–3. We found that the SMGs for which the mid-IR emission is dominated by star formation orAGN have extended millimeter-sizes, with respective median R c , e = 1 . +0 . − . and 1.5 +0 . − . kpc. Instead, the SMGs forwhich the mid-IR emission corresponds to star-forming/AGN composites have more compact millimeter-wave sizes,with median R c , e = 1 . +0 . − . kpc. The relation between millimeter-wave size and AGN fraction suggests that this sizemay be related to the evolutionary stage of the SMG. The very compact sizes for composite star-forming/AGN systemscould be explained by supermassive black holes growing rapidly during the SMG coalescing, star-formation phase. Keywords: submillimeter: galaxies — galaxies: evolution — galaxies: formation — galaxies: high-redshift
Corresponding author: Soh [email protected]
Ikarashi et al. INTRODUCTIONThe morphology and size of star-forming regions insubmillimeter galaxies (SMGs) are important propertieswith which we can address the nature of their prodi-gious, dust-obscured star formation, and consequentlythe formation and evolution of the most massive galax-ies. The Atacama Large Millimeter/submillimeter Ar-ray (ALMA) is enabling astronomers to image high-redshift SMGs with angular resolutions of . ′′ .3. SomeALMA studies have reported effective radii ( R e ) of ∼ . z = 1–3, as determined from mid-IRdata. We adopt throughout a cosmology with H =70 km s − Mpc − , Ω M = 0 .
3, and Ω Λ = 0 . ALMA OBSERVATIONS AND SAMPLESThe sample used in this paper comes from ourALMA 1100- µ m continuum imaging survey of 144 brightAzTEC/ASTE sources with F µ m , AzTEC ≥ . XMM-Newton
Deep Field (SXDF;Furusawa et al. 2008). The SXDF survey was con-ducted in the ALMA Cycles 2 and 3 (2013.1.00781,2015.1.00442.S: PI. Hatsukade; B. Hatsukade et al. 2017,in preparation). The ALMA observations in Cycle 2 were carried outwith the array configurations C34-5 and C34-7, with 37–38 working 12-m antennas covering up to a uv distanceof ∼ λ . In Cycle 3, the observations were ex-ecuted in array configuration C40-4, covering up to a uv distance of ∼ λ . On-source integration timesper source in each cycle were 0.6 min. The typical syn-thesized beam size for our ALMA continuum imagesis ∼ . ′′ × . ′′
23 (PA ∼ ◦ ), after combining theCycle 2 and 3 data. The average r.m.s. noise levelis 120 µ Jy beam − . The images were generated withBriggs weighting, using a robust parameter of 0.3.The ALMA continuum maps yielded 70 ALMA-identified AzTEC SMGs (hereafter ASXDF SMGs) with S peak /N ≥
10 detections, suitable for reliable ALMAmillimeter-wave size measurements (e.g. Ikarashi et al.2015). We removed one lensed SMG (ASXDF1100.001;Ikarashi et al. 2011), leaving 69 SMGs. ALMA fluxeswere re-measured in tapered ALMA images with a syn-thesized beam of ∼ ′′ .
6, which is larger than the mea-sured mm-wave sizes of SMGs in this paper, using theIMFIT task in CASA.For 51 ASXDF SMGs, we obtained well-constrainedphotometric redshifts, with a median error δz = 0 . ± .
02, based on the individual 1- σ errors estimated by LePhare (e.g. Ilbert et al. 2006) in spectral energy distri-bution (SED) model fitting using the B , V , Rc , i ′ , z ′ , J , H , Ks , 3.6 and 4.5 µ m data (S. Ikarashi et al. 2017,in preparation). The remaining SMGs lie outside thecoverage of the optical/near-IR images, or have individ-ual 1- σ errors of >
1. Photometric and spectroscopicredshifts from the literature are listed in Table 1. ALMA MILLIMETER-WAVE SOURCE SIZEMEASUREMENTSWe measured millimeter-wave sizes as circularized ef-fective radii ( R c , e ) for the 69 ASXDF SMGs with ALMAvisibility data, in the same manner as Ikarashi et al.(2015). We used uv -distance versus amplitude plots(hereafter uv -amp plots) for our measurements. Al-though the ALMA data cover uv distances up to ∼ λ , we used only data at ≤
500 k λ , which corre-sponds to a scale of ∼ . ′′
2. Adopting this cutoff forthe longest uv distance is the equivalent of smoothingwith a larger size kernel in the image plane. We aim tomitigate the effects of possible clumpy structures in thesize measurements and to measure R c , e robustly. Forthe sources detected with ≥ σ in the ALMA Cycle-2images alone, we measured their sizes using only Cycle-2 data, to avoid effects due to any systematic absoluteflux calibration offsets between our Cycle 2 and 3 data . We measured sizes by fitting a Gaussian model to theobserved data in the uv -amp plots. When we measure Comparisons of the fluxes of ASXDF sources in our Cycle-1, 2 and 3 data indicated that the fluxes in the Cycle-3 data are ery compact millimeter sizes for composite star-forming/AGN SMGs ≥ σ ) in each ALMA imagewere removed from the visibility data based on simplesource properties derived by IMFIT task.In order to estimate possible systematics in the sizemeasurements, we injected mock sources into ALMAnoise visibility images, generated from the actual ALMAdata as in Ikarashi et al. (2015). Briefly we injected asymmetric Gaussian component for a range of sourcesizes and flux densities that cover the putative param-eter range of our ASXDF sources with uniform proba-bility. As tested in Ikarashi et al. (2015), our methodcan measure circularized effective radii correctly evenif a source has an asymmetric Gaussian profile. Wecorrected our raw measured sizes based on the resultsof the simulations for the data used in this paper. Asa result, we obtained ALMA millimeter-wave sizes of0 ′′ . ′′ .
68 (FWHM) for the 69 ASXDF SMGs. Notethat ASXDF1100.009.1 has two distinct millimeter-wavecomponents with a separation of ∼ ′′ .6, sharing a hostgalaxy at z spec = 0 . RELATION BETWEEN MILLIMETER SIZESAND FLUXESFig. 1 (left panel) shows all 69 ASXDF SMGs in anALMA 1100- µ m vs. millimeter-wave size plot. Addi-tionally, we plot 13 ALMA-identified, fainter SMGs at z & ≥ σ continuum detection based on simple Gaussian modelsexplains the absence of SMGs in the top-left corner. Thebottom-right corner, instead, is free from any such se-lection biases, so the absence of SMGs requires an ex-planation.The absence of SMGs in the bottom-right corner ofFig. 1 can be interpreted as the influence of Eddington-limited star formation (Murray et al. 2005). Accord-ing to Younger et al. (2008), which reported pioneeringstudies of maximum star formation in bright SMGs, amaximum star-formation rate is given by SF R max = 480 σ D kpc κ − M ⊙ yr − , (1)where D kpc is the characteristic physical scale of thestarburst region in kpc, σ is the line-of-sight gasvelocity dispersion in units of 400 km s − , and κ isthe dust opacity in units of 100 cm g − . Here weadopt a Chabrier initial mass function (Chabrier 2003); κ = 1, as in Younger et al. (2008); and a mediangas velocity dispersion of 510 km s − from CO line ob-servations of SCUBA SMGs (Bothwell et al. 2013). Wealso adopt 2 × FWHM or 4 × R c , e , which is expected toinclude 94% of the total light, as D kpc . The derived SF R max was corrected with this factor of 0.94. systematically ∼
20% smaller. Therefore, we corrected the primaryflux calibration for this effect.
In order to plot the relation between SFR and phys-ical scale described by Equation 1 on Fig. 1 (the leftpanel), we assume a fixed redshift z = 2. The conver-sion factors from ALMA fluxes to SFRs were derivedby bootstrapping given a dust temperature ( T d ) distri-bution for lensed 1.3mm-selected galaxies (Weiß et al.2013) and an SED library with T d information compiledin Swinbank et al. (2014). For these assumptions, weobtain a possible range of Eddington-limited star for-mation rates.For a more direct comparison of the millimeter fluxesand sizes of SMGs with Eddington-limited star forma-tion, we re-plot 51 of the 69 SMGs at z = 0 . F µ m , given the range ofpossible dust temperatures T d and SEDs noted above.We assume that the AGN contribution to the submil-limeter flux is negligible (see references in Rosario et al.2012). In order to visualize the coverage of the size-SFRplane produced by the large SFR uncertainties (due tothe unknown dust SED temperatures), we show the fullSFR probability density distribution (rather than a sin-gle value) for each SMG. The results in both panels ofFig. 1 show that the SMGs avoid the SFR region aroundthe Eddington limit, suggesting that the minimum pos-sible millimeter-wave sizes for bright SMGs are given bythe Eddington limited star formation.The empirical relation between flux and size can ex-plain the apparent discrepancy between the reported(sub)millimeter-wave (median) sizes of 0 . ′′ +0 . ′′ − . ′′ byIkarashi et al. (2015) and 0 . ′′ ± . ′′
04 by Simpson et al.(2015). Given F µ m /F µ m = 2 for conversionof the observed fluxes, Simpson et al. (2015) covered F µ m & . . ′′ +0 . ′′ − . ′′ . RELATION BETWEEN MILLIMETER SIZESAND AGNWe present our studies of the connection betweenthe millimeter-wave sizes and AGN in SMGs, basedon a mid-IR AGN diagnostic. We consider 25 ALMA-identified SMGs with 1 < z phot or spec <
3, which are de-tected in all IRAC and MIPS 24 µ m images. All SMGshere have redshift information and a single componentat ∼ ′′ .2 resolution. More than 15 out of the 25 arelocated above 4 × the main sequence at z ∼ z = 1–3,four are not considered in our analysis: two SMGs arenot detected at 24 µ m and the other two are blended inthe IRAC maps.5.1. Mid-IR AGN diagnostic
A 4.5 µ m/8 µ m/24 µ m color-color plot has often beenused as an AGN diagnostic for high-redshift, dusty Ikarashi et al.
Figure 1.
Left:
ALMA 1100- µ m flux vs. ALMA millimeter-wave size for the ASXDF SMGs with and without redshiftinformation (filled black and blue circles, respectively). The black points correspond to the ASXDF SMGs obtained in ourALMA Cycle-2 and 3 projects, as analyzed in this paper. The grey shaded area shows the approximate source selection limit(10 σ ) on our ALMA images. The orange points show other ASXDF SMGs at z & σ ranges of T d and gas velocity dispersion of known SMGsfrom the literature. The red solid line shows the Eddington-limited star-formation relation for the median T d and gas velocitydispersion. Right:
SFR vs. effective radii in physical scale for the 51 ASXDF SMGs with available photometric or spectroscopicredshifts. The selection limit assumes a physical scale for z = 2. The background grey-shaded area shows P ( SF R, size ) foreach SMG, taking into account the large uncertainty of the SFR due to the unknown dust temperature T d . The Eddington-limitrelation is indicated with magenta lines (solid for the median and dashed for ± σ of the gas velosity dispersion). Typical errorbars are indicated with a red cross in the upper part of the plot. Open circles in both panels mark ASXDF1100.009.1, whichhas two distinct components in the ALMA image. infrared-luminous galaxies, such as SMGs and DOGs at z ∼ z galaxies. Empirical SEDtemplates (top left panel in Fig. 3) suggest that high-redshift galaxies dominated by star formation or AGNin mid-IR light can be segregated from each other inthe mid-IR color-color plane. The position of our 25SMGs in this color-color plot shows that some of themdo not follow either the model tracks for star-formation-dominated or AGN-dominated galaxies.We generated the expected mid-IR colors of galax-ies that are a composite of SF and AGN by combiningSEDs of SF and AGN with various SF/AGN ratios. This‘toy’ color prediction reproduces the colors of ‘compos-ite SMGs’ which are likely to be dominated by neitheran AGN nor a starburst in the mid-IR (top right panelin Fig. 3).We divided the 25 SMGs into four sub-groups basedon their 4.5/8/24- µ m colors: star-forming, composite,AGN-dominant and ‘no class’. The criteria are: • F µ m /F . µ m < . V F µ m /F µ m ≥ • F µ m /F . µ m ≥ . V F µ m /F µ m ≥ • F µ m /F . µ m ≥ . V F µ m /F µ m < • F µ m /F . µ m < . V F µ m /F µ m < F . µ m /F . µ m =1.15 as criterion for separation, as thisensures that all galaxies that satisfy neither an AGNcriteria by Donley et al. (2012) nor another criteria byStern et al. (2005) also lie on the star-forming region ofthe colour-colour diagram. The predicted 24 µ m/8 µ mcolor evolution with redshift, as derived by public em-pirical mid-IR SED templates for high- z star-forming ery compact millimeter sizes for composite star-forming/AGN SMGs Figure 2.
Stellar mass vs. SFR for the 25 ASXDF SMGsat z ∼ . ≤ z ≤ . § galaxies, composite galaxies, and AGN dominant galax-ies (Kirkpatrick et al. 2015), are shown along with oursample SMGs (bottom left, Fig. 3). For these templates,the respective mid-IR AGN fractions of each sample are <
20, 20–80, and ≥ λ rest = 8 µ m. The predictions based on the Kirk-patrick et al. SED templates suggest that our criteria for24 µ m/8 µ m color can work to select an AGN-dominantclass, and show that our composite type is expected tohave typically AGN fractions of around ∼ Results
In the millimeter-wave physical size vs. SFR plot (bot-tom right panel in Fig. 3), all SMGs with compositemid-IR components are evidently more compact and lo-cated closer to the Eddington limit than the other SMGswith star-forming dominant or AGN dominant mid-IRcomponents.The respective median R c , e for the SMGs classifiedas star-forming dominant, composites, and AGN dom-inant are 1.6 +0 . − . , 1.0 +0 . − . , and 1.5 +0 . − . kpc. The sizedifference between the SMGs with composite and star-forming mid-IR components, and the difference betweenthe SMGs with composite and AGN-dominant mid-IRcomponents are real, with a significance level of > XMM-Newton
X-ray maps(Ueda et al. 2008), probably because these maps aretoo shallow. Nevertheless, we can compare our resultswith the sizes derived for the host galaxies of five high- z ,X-ray-selected AGN ( L − = 10 . − . erg s − ) byHarrison et al. (2016). These authors reported a sizedistribution for their AGN hosts similar to the SMGsizes in Simpson et al. (2015). The most X-ray luminoussource in their sample (with L − = 10 . erg s − )has an extended size, and the remaining four ( L − =10 . − . erg s − ) have compact sizes, which are com-parable to those of our composite type here (Fig. 3,bottom right).5.3. AGN growth during a very compact star-formingphase?
The very compact millimeter-wave sizes of the SMGswith composite mid-IR components suggest that a cen-tral supermassive black hole could be growing in a com-pact and coalescing star-forming phase, which is con-sistent with the predictions of Springel et al. (2005) forgalaxy major mergers. The extended millimeter-wavesizes of the SMGs of the star-forming dominant classcan be explained by a mid-stage merger as seen in, e.g.,VV114 (Saito et al. 2015). Actually ASXDF1100.055.1with the star-forming dominant class shows merger-like near-IR morphology (Fig. 4). Instead, the extendedsizes of the SMGs with the AGN-dominant class are puz-zling. In line with the evolutionary scenarios of, e.g.,Sanders et al. (1988); Hopkins et al. (2008); Toft et al.(2014) that SMGs evolve into QSOs, these extendedsizes may be explained by positive AGN feedback bya growing supermassive black hole in the phase ofstar-formation quenching, as it is suggested by simula-tions for luminous AGN/QSOs (e.g. Ishibashi & Fabian2012; Zubovas et al. 2013) and considered for someluminous QSOs (e.g. Carniani et al. 2016). In fact,ASXDF1100.057.1 with the AGN dominant class hasa QSO-like near-IR morphology (Fig. 4). However, nosignificant near-IR morphological difference betweenAGN-host and non-AGN-host galaxies, that are notsubmillimeter selected, is reported (e.g. Kocevski et al.2012). The extended submillimeter sizes in our SMGsmay come from the nature of their host galaxies.
Facilities:
ALMA,Spitzer,Subaru,UKIRT,HST(STIS)
Ikarashi et al.
Figure 3.
Relations between ALMA millimeter-wave size, SFR, and mid-IR color.
Top:
IRAC 4.5, 8, and MIPS 24- µ m-colorAGN diagnostic for z ∼ Top left:
The colored shaded areas mark the diagnosticcriteria of star-formation (SF) dominant, composite and AGN-dominant in mid-IR light. The solid curves are the predictionsbased on the SEDs in the SWIRE Template Library (Polletta et al. 2007), which is mainly composed of local star-forminggalaxies, (U)LIRGs, Seyfert galaxies, and QSOs. The colored filled circles indicate the ALMA-identified SMGs.
Top right:
Simulated mid-IR colors of mock galaxies based on empirical SED templates, with the color points showing the AGN fractionbased on the mock 8 µ m fluxes. The black points correspond to the ASXDF SMGs. The dashed lines show the criteria forSF/AGN classification. Bottom left:
Redshift versus 24- µ m/8- µ m colors for our sample. The solid lines indicate color evolutionpredictions based on empirical SED templates derived from star-forming-dominant, composite and AGN-dominant high- z galaxytemplates from Kirkpatrick et al. (2015). The dashed red line corresponds to an AGN-dominated system based on a local QSOSED template in the SWIRE template library. Open squares and circles indicate SMGs that satisfy the Donley et al. and Sternet al. IRAC AGN criteria, respectively. Bottom right:
SFR vs. millimeter-wave effective radius. The colored dotted lines delimitareas where P ( size, SF R ) > = 0 . × P peak for each SF/AGN type. Host galaxies of X-ray-selected AGN from Harrison et al.(2016) (H+16) are marked by red stars. The size distribution of our SMGs is shown in the histogram on the right-hand side ofthe plot. ery compact millimeter sizes for composite star-forming/AGN SMGs Figure 4.
HST images with ALMA continuum contours for three of our galaxies. All images are R(1.6 µ m)/G(1.2 µ m)/B(0.8 µ m)composites from CANDELS-UDS. The cyan contours correspond to the ALMA 1100- µ m continuum (3, 5 and 7 σ ;1 σ ∼ µ Jy beam − ). The magenta circles indicate the circularized effective radii of ALMA 1100- µ m continuum emission. Thegreen circles correspond to the 1.6- µ m continuum. The respective effective radii ( R nir , e ) are 6.3 ± ± × − , and3.6 ± HST sizes are based on van der Wel et al.(2012).
Ikarashi et al.
Table 1.
Summary data of the ASXDF SMG sample analyzed in this paper.
ID R.A. Dec. SNR F µ m z photo SFR mm-wave size mm-wave size AGN(J2000) (J2000) (mJy) ( M ⊙ yr −
1) (FWHM; arcsec) ( R c , e; kpc) (mid-IR)ASXDF1100.002.1 2:17:30.63 -4:59:36.8 15 4.81 ± . − .
87 990+720 −
340 0.42+0 . − .
02 1.6+0 . − . · · · ASXDF1100.004.1 2:18:05.65 -5:10:49.7 14 4.39 ± . − .
16 880+420 −
290 0.40+0 . − .
04 1.5+0 . − . · · · ASXDF1100.005.1 2:17:30.45 -5:19:22.5 25 7.24 ± . − .
01 1200+990 −
420 0.34+0 . − .
02 1.2+0 . − . · · · ASXDF1100.006.1 2:17:27.32 -5:06:42.8 10 5.11 ± . − .
15 930+340 −
330 0.68+0 . − .
06 2.2+0 . − . · · · ASXDF1100.007.1 2:18:03.01 -5:28:42.0 20 6.26 ± . − .
22 1300+930 −
450 0.32+0 . − .
02 1.2+0 . − . · · · ASXDF1100.008.1 2:16:47.93 -5:01:29.9 12 6.45 ± . − .
08 1500+950 −
460 0.62+0 . − .
06 2.6+0 . − . ± a −
190 0.30+0 . − .
04 0.9+0 . − . · · · ASXDF1100.009.1B 2:17:42.16 -4:56:28.5 11 1.16 ± a −
50 0.10+0 . − .
06 0.6+0 . − . · · · ASXDF1100.011.1 2:17:50.59 -5:30:59.2 13 4.22 ± . − .
63 730+440 −
260 0.38+0 . − .
04 1.1+0 . − . · · · ASXDF1100.014.1 † ± . − .
03 690+270 −
210 0.50+0 . − .
08 2.1+0 . − . ± . − .
06 850+390 −
240 0.24+0 . − .
04 0.8+0 . − . · · · ASXDF1100.018.1 2:18:13.83 -4:57:43.5 14 3.47 ± . − .
02 850+650 −
280 0.26+0 . − .
04 1.1+0 . − . • ± . − .
01 1100+460 −
380 0.30+0 . − .
02 1.2+0 . − . · · · ASXDF1100.021.1 2:18:16.49 -4:55:08.8 16 4.03 ± . − .
04 920+720 −
310 0.28+0 . − .
04 1.1+0 . − . ± . − .
06 710+550 −
240 0.20+0 . − .
04 0.8+0 . − . ± . − .
12 480+350 −
160 0.16+0 . − .
06 0.6+0 . − . · · · ASXDF1100.025.2 † ± . − .
07 470+320 −
150 0.34+0 . − .
04 1.3+0 . − . · · · ASXDF1100.029.1 † ± . − .
17 570+360 −
180 0.46+0 . − .
10 1.8+0 . − . † ± . − .
12 480+380 −
170 0.28+0 . − .
06 1.1+0 . − . ± c −
350 0.34+0 . − .
02 1.4+0 . − . ± b −
220 0.16+0 . − .
06 0.7+0 . − . · · · ASXDF1100.035.1 † , • ± . − .
11 450+360 −
150 0.52+0 . − .
08 2.1+0 . − . · · · ASXDF1100.041.1 2:17:53.87 -5:26:35.7 10 2.91 ± . − .
00 520+260 −
180 0.42+0 . − .
10 1.6+0 . − . · · · ASXDF1100.042.1 2:18:38.29 -5:03:18.3 12 3.26 ± . − .
01 680+440 −
240 0.42+0 . − .
06 1.6+0 . − . · · · ASXDF1100.044.1 2:17:45.85 -5:00:56.7 12 1.93 ± . − .
72 330+210 −
84 0.09+0 . − .
05 0.2+0 . − . · · · ASXDF1100.046.1 2:17:13.34 -4:58:57.4 16 4.00 ± . − .
10 810+620 −
280 0.28+0 . − .
04 1.0+0 . − . · · · ASXDF1100.047.1 † ± . − .
02 500+400 −
160 0.40+0 . − .
06 1.6+0 . − . † ± . − .
12 570+460 −
200 0.40+0 . − .
04 1.6+0 . − . ⋆ ± . − .
15 700+360 −
240 0.24+0 . − .
08 0.9+0 . − . · · · ASXDF1100.051.1 † ± . − .
04 430+270 −
150 0.08+0 . − .
04 0.3+0 . − . · · · ASXDF1100.051.2 † ± . − .
15 520+270 −
160 0.30+0 . − .
06 1.0+0 . − . · · · ASXDF1100.052.1 † ± . − .
65 440+340 −
150 0.34+0 . − .
06 1.3+0 . − . † ± . − .
24 570+290 −
180 0.34+0 . − .
06 1.4+0 . − . ± . − .
11 820+360 −
260 0.34+0 . − .
06 1.4+0 . − . ± . − .
41 750+550 −
230 0.34+0 . − .
06 1.1+0 . − . · · · ASXDF1100.077.1 † ± . − .
12 320+190 −
110 0.22+0 . − .
08 0.8+0 . − . · · · ASXDF1100.089.1 2:18:10.64 -5:34:53.6 21 4.73 ± . − .
09 830+600 −
200 0.24+0 . − .
02 0.7+0 . − . · · · ASXDF1100.095.1 † ± . − .
08 440+320 −
150 0.32+0 . − .
08 1.3+0 . − . ± . − .
08 670+550 −
210 0.24+0 . − .
04 1.0+0 . − . ± b −
220 0.24+0 . − .
08 1.0+0 . − . † ± . − .
86 310+190 −
80 0.34+0 . − .
06 1.1+0 . − . · · · ASXDF1100.115.1 2:16:59.42 -5:10:55.8 12 4.23 ± a −
220 0.50+0 . − .
06 1.7+0 . − . · · · ASXDF1100.134.1 2:17:54.80 -5:23:23.8 15 3.27 ± . − .
05 740+500 −
260 0.24+0 . − .
04 1.0+0 . − . ± . − .
10 810+630 −
260 0.34+0 . − .
06 1.4+0 . − . † ,⋆ ± . − .
20 530+450 −
180 0.22+0 . − .
08 0.9+0 . − . · · · ASXDF1100.203.1 † ± . − .
15 440+330 −
150 0.34+0 . − .
10 1.4+0 . − . ± . − .
14 1400+760 −
510 0.34+0 . − .
02 1.2+0 . − . · · · ASXDF1100.228.1 2:18:09.66 -5:18:43.1 12 3.11 ± . − .
14 740+610 −
240 0.38+0 . − .
06 1.6+0 . − . ± . − .
11 820+620 −
270 0.26+0 . − .
08 1.1+0 . − . ± . − .
14 1100+820 −
370 0.26+0 . − .
04 1.1+0 . − . † ± . − .
02 370+250 −
120 0.15+0 . − .
09 0.6+0 . − . † ± . − .
14 410+260 −
140 0.24+0 . − .
10 1.0+0 . − . † ± · · · · · · . − . · · · · · · ASXDF1100.010.1 2:17:39.79 -5:29:19.2 24 5.94 ± · · · · · · . − . · · · · · · ASXDF1100.026.1 † ± · · · · · · . − . · · · · · · ASXDF1100.040.1 2:17:55.24 -5:06:45.1 15 3.14 ± · · · · · · . − . · · · · · · ASXDF1100.053.1 2:16:48.20 -4:58:59.6 10 4.02 ± · · · · · · . − . · · · · · · ASXDF1100.054.1 2:17:15.41 -4:57:55.6 11 4.12 ± · · · · · · . − . · · · · · · ASXDF1100.068.1 2:17:42.17 -5:25:46.8 12 3.24 ± · · · · · · . − . · · · · · · ASXDF1100.070.1 † ± · · · · · · . − . · · · · · · ASXDF1100.074.1 2:18:33.31 -4:58:07.0 10 2.77 ± · · · · · · . − . · · · · · · ASXDF1100.097.1 2:18:18.54 -5:34:34.7 11 2.53 ± · · · · · · . − . · · · · · · ASXDF1100.097.2 † ± · · · · · · . − . · · · · · · ASXDF1100.133.1 2:18:05.51 -5:35:46.5 11 2.25 ± · · · · · · . − . · · · · · · ASXDF1100.161.1 † ± · · · · · · . − . · · · · · · ASXDF1100.168.1 2:18:04.37 -5:34:03.5 11 1.79 ± · · · · · · . − . · · · · · · ASXDF1100.213.1 † ± · · · · · · . − . · · · · · · ASXDF1100.231.1 2:17:59.65 -4:46:49.8 12 2.88 ± · · · · · · . − . · · · · · · ASXDF1100.243.1 † ± · · · · · · . − . · · · · · · ASXDF1100.252.1 2:17:05.65 -5:15:04.9 12 2.62 ± · · · · · · . − . · · · · · · Notes. † ALMA flux, SNR, and size measurements are conducted in the ALMA data after combining the Cycle 2 and 3 data.For sources without † , all ALMA measurements were done in the ALMA Cycle-2 data. ⋆ The SMGs are not included in the analysis in § µ m. • The SMGs are not included in the analysis in § § a spectroscopic redshifts by cross-identification with the UDS-z survey catalog (e.g. Bradshaw et al. 2013; McLure et al. 2013). b spectroscopic redshifts by cross-identification with the SCUBA SMGs (Banerji et al. 2011). c spectroscopic redshifts by cross-identification with the SCUBA SMGs (Coppin et al. 2010). ery compact millimeter sizes for composite star-forming/AGN SMGs9REFERENCES