Unveiling Sizes of Compact AGN Hosts with ALMA

The processes driving the coevolution of galaxies and their supermassive black holes remain a largely debated issue in extragalactic astrophysics. Specifically, the connection between the formation of stars in galaxies and the fueling of their central black holes is still not fully understood. Previous results revealed that X-ray and optically selected active galactic nuclei (AGN) reside in galaxies harboring sustained activity of star formation (e.g., Harrison et al. 2012; Mullaney et al. 2012; Juneau et al. 2013; Rosario et al. 2013). Furthermore, although high luminosity AGNs seem to be connected to violent events (i.e., quasars) and may reside in galaxies more often involved in major mergers (e.g., Treister et al. 2009; Kartaltepe et al. 2012), the majority of moderate-luminosity X-ray selected AGNs appear to live mostly in disk-dominated isolated systems (Mainieri et al. 2011; Silverman et al. 2011; Fan et al. 2014; Villforth et al. 2014). Also, major mergers do not appear to be the dominating triggering mechanism for luminous unobscured AGN (Mechtley et al. 2016; Villforth et al. 2017). Their morphology should thus be very similar to that of sources populating the so-called main sequence (stellar mass versus star formation rate (SFR); e.g., Brinchmann et al. 2004; Daddi et al. 2007; Elbaz et al. 2007; Noeske et al. 2007; Ilbert et al. 2015; Schreiber et al. 2015; Chang et al. 2015) of star-forming galaxies (SFGs; e.g., Gabor et al. 2009; Cisternas et al. 2011; Kocevski et al. 2012), implying that the bulk of supermassive black hole accretion is likely driven by internal processes and not by major mergers. Nonetheless, it is still an open question whether the internal structure of AGN hosts is comparable to or different from that of normal SFGs.

Most current studies have focused on samples of SFGs. Recent observations from the Atacama Large Millimeter/submillimeter Array (ALMA) allow us to measure far-infrared (FIR) sizes with high angular resolution and sensitivity (e.g., Ikarashi et al. 2015; Simpson et al. 2015a, 2015b; Harrison et al. 2016; Hodge et al. 2016). For size comparison, it has been discussed that rest-frame FIR sizes of massive SFGs are smaller than their rest-frame optical sizes (e.g., Chen et al. 2015; Barro et al. 2016; Fujimoto et al. 2017, 2018; Tadaki et al. 2017; Elbaz et al. 2018; Lang et al. 2019). A possible explanation of the central compact component is a compact dusty bulge which is related to bulge formation in the center (Fujimoto et al. 2017). Moreover, it has been discussed that massive galaxies turn out to exhibit compact star formation (Barro et al. 2016; Elbaz et al. 2018), and transform into compact quiescent galaxies at z ∼ 2 (e.g., van der Wel et al. 2014). There have been several attempts to estimate FIR sizes for faint submillimeter galaxies (Rujopakarn et al. 2016; Fujimoto et al. 2017; González-López et al. 2017), but the uncertainties are still constrained by the small number statistics. Moreover, it has been discussed that both high sensitivities and high angular resolutions are required to measure FIR sizes for less massive objects (Fujimoto et al. 2018). On the other hand, it is still unclear whether the existence of AGNs affects the structures of their host galaxies, and whether AGN host galaxies show different rest-frame FIR sizes compared to other SFGs.

Chang et al. (2017b) investigated AGN and non-AGN host morphologies through Hubble Space Telescope (HST)/F814W images, and found that obscured IR-selected AGN hosts are more compact than normal star-forming sources at a given stellar mass at z ∼ 1. Following that work, in this paper we resolve the dust emission of AGN host galaxies using high-resolution ALMA imaging. We show that their compact sizes are consistent with what we found using HST/F814W-band imaging, and the FIR (ALMA/ 1 mm) sizes are yet even smaller than the optical sizes. We compare all available FIR sizes of AGN and non-AGN hosts in the Cosmic Evolution Survey (COSMOS; Scoville et al. 2007) field. For comparison, we also measure the FIR sizes of a control sample of non-AGN SFGs in the COSMOS field for which we find complementary high-resolution ALMA imaging in the archive. The structure of this paper is as follows. We describe the data and sample selections in Section 2. We analyze the physical and structural properties in Section 3. We discuss the results in Section 4 and summarize in Section 5. Throughout the paper, we use AB magnitudes, adopt the cosmological parameters (Ω_M, Ω_Λ, h) = (0.30, 0.70, 0.70), and assume the stellar initial mass function of Chabrier (2003).

2.1. Our ALMA Observations

Our ALMA Band 7 (275–373 GHz) observations (#2016.1.01355.S) contain 10 obscured AGN host galaxies selected in the infrared (based on identifying mid-IR power-law signatures which are characteristic of AGNs as opposed to star-forming dominated sources; more details in Chang et al. 2017a) at 0.5 < z < 1.0 in the COSMOS field. Our targets were selected along the main sequence within 0.3 dex of star formation, as well as a star-forming sample in the rest-frame UVJ plot (Williams et al. 2009), with stellar masses $\mathrm{log}\,{M}_{* }\gt 10.5$ in a 24 μm selected sample in the COSMOS field. Reliable spectroscopic redshifts are available. In order to ensure a robust estimate of the dust-obscured star formation, we restricted our sample to Herschel-detected sources. We confirmed that these objects show compact F814W-band radial profiles as described in Chang et al. (2017b).

We proposed the standard set-up for single continuum centered at 343.5 GHz of Band 7 over a 7.5 GHz bandwidth. The proposed angular resolution is 01, which is close to the resolution of the F814W-band imaging. The array configuration was C40-6 with 40 antennas, and the proposed rms value is 0.1 mJy/beam per 7.5 GHz. The data were reduced with the Common Astronomy Software Applications package (CASA; McMullin et al. 2007), and the steps of data reduction and calibration were performed with the scripts provided by the ALMA observatory. The entire map was produced by the CLEAN algorithm with the TCLEAN task. The final images are characterized by a synthesized beam size of 012 × 010¹² as shown in Table 3. One of the objects (Source name: 972308) is detected, and is shown in Figure 1 along with the remaining detections from the ALMA archive. The rms value is 0.096 mJy/beam per 8 GHz bandwidth (∼0.099 mJy/beam per 7.5 GHz). For the remaining nine objects, we stacked their imaging and computed the median values in each pixel to create a stacked imaging as shown in Figure 2. The peak flux is 0.09 mJy ± 0.03 mJy (3.39σ; rms value is calculated from the stacked image).¹³ We also stacked these objects by STACKER (Lindroos et al. 2015) in the uv plane. The peak flux is 0.08 mJy ± 0.03 mJy (2.24σ; rms value is calculated from the output image of STACKER), which is consistent with the stacking results from the image plane. For point sources, nine objects with an rms of 0.1 mJy/beam can produce at best a stacked image with an rms of 0.03 mJy/beam, and given that the stacked signal is 0.09 the signal-to-noise ratio (S/N) would be around 3. Our stacking result from the image plane shows slightly large signal-to-noise ratios, which could be explained by the extended profile and the beam size. It is not straightforward to convert a stacked image in the uv plane from angular scale to physical scale, so we adopted the FIR stacked image in the image plane in this paper.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

2.2. Complementary Sample from the Archive

2.3. Spectral Energy Distribution (SED) Fitting Results

2.4. Size Measurements

We separated the sample between non-AGN SFGs, X-ray selected AGNs (XAGNs: L_X(2–10 keV) > 10⁴² ergs/s), and infrared-selected AGNs (IRAGNs: using the mid-infrared colors from the IRAC 3.6, 5.8, 4.5, and 8.0 μm; obscured IRAGNS: ${L}_{\mathrm{IR},\mathrm{AGN}}/{L}_{{\rm{X}},\mathrm{AGN}}\gt 20$), that are identified by previous work (Civano et al. 2016; Marchesi et al. 2016; Chang et al. 2017a, 2017b). Figure 3 shows the SFR versus stellar mass diagrams in the two redshift bins. Most of the sources in the sample from the ALMA archive are massive and have high SFRs, and they are in general on and above the star-forming sequence. Our star-forming sequences are estimated by the SED fitting results for $\mathrm{NUV}\mbox{--}r$ versus r−J selected galaxies. We also show the main-sequence relations from Speagle et al. (2014) and Whitaker et al. (2014). Our SFR estimations are generally lower than those in the literature by ∼0.1–0.3 dex at z ∼ 2. The main reason might be selection of the star-forming sample. With the same sample selection, our main sequences are very close (<0.1 dex) to the main sequences calculated by the stellar masses and SFRs in the COSMOS2015 catalogs. To avoid bias from different selection criteria, we use the $\mathrm{NUV}-r$ versus r − J selection, but show the literature in Figure 3 for comparison. The SFR differences between the various works are more significant at the high stellar mass end, so it is important to compare AGN hosts with non-AGN galaxies using SFRs derived with the same method. Comparing to the archival data at z ∼ 1, our AGN sample have low SFRs because they were selected to be within ±0.3 dex of the main sequence. This may be the reason why our sample was mostly undetected with ALMA. The only one IRAGN at z ∼ 1 has lower stellar mass (∼0.14 dex) and higher SFR (∼0.20 dex) than non-AGN galaxies. There is only one IRAGNs at z ∼ 2, with lower stellar masses (∼0.04 dex) and lower SFRs (∼0.35 dex) than the median of the non-AGN galaxies.

Zoom In Zoom Out Reset image size

Figure 4 shows the FIR size as a function of redshift (left), stellar mass (middle), and the distance to the star-forming sequence (right) as shown in Figure 3. The sizes of our detected (0.43 ± 0.28 kpc) and stacked (0.90 ± 0.21 kpc) obscured AGN hosts are very compact compared to those from the archive data (mean = 1.76 ± 0.25 kpc and median = 1.67 ± 0.30 kpc at z ∼ 1 and $\mathrm{log}\,{M}_{* }\sim 11$). The uncertainties here and below are estimated from a bootstrapping analysis. As shown in Table 1, for $\mathrm{log}\,{M}_{* }\sim 11$ sources at z ∼ 1, the median (mean) FIR sizes of the SFGs, XAGNs, IRAGNs, are 1.74 ± 0.50 (1.62 ± 0.50) 2.46 (one object),¹⁵ 0.77 ± 0.50 (0.76 ± 0.48), respectively. As shown in Table 1, for $\mathrm{log}\,{M}_{* }\sim 11$ sources at z ∼ 2, the median (mean) FIR sizes of the SFGs, XAGNs, and IRAGNs, are 1.25 ± 0.73 (1.30 ± 0.39), 0.69 ± 0.25 (0.70 ± 0.24), 0.70 (one object), respectively. We find systematically smaller galaxy sizes at z ∼ 2 relative to the corresponding samples at z ∼ 1. The measurements of HST/F814W images are rest-frame optical sizes for z ∼ 1 sources but rest-frame UV sizes for z ∼ 2 sources (e.g., van der Wel et al. 2014), and color gradient which shows redder centers should also be taken into account (e.g., Wuyts et al. 2012). Most importantly, the number of objects in each sample is limited, especially for the AGN sample. Our IRAGN sample was selected to be within ±0.3 dex of the main sequence, so we also consider $\mathrm{log}\,{M}_{* }\sim 11$ SFGs on the main sequence. For $\mathrm{log}\,{M}_{* }\sim 11$ SFGs on the main sequence (the distance to the star-forming sequence <0.5 dex; a less strict range because of the limited sample size) at z ∼ 1, the median (mean) sizes are 1.97 ± 0.44 kpc (1.83 ± 0.49 kpc). For $\mathrm{log}\,{M}_{* }\sim 11$ SFGs on the main sequence (the distance to the star-forming sequence <0.5 dex) at z ∼ 2, the median (mean) sizes are 1.26 ± 1.43 kpc (1.27 ± 1.70 kpc). As a result, we find that IRAGN hosts are in general smaller than SFGs on the main sequence at a given stellar mass and redshift bin. However, it is difficult to find a reasonable parent sample for our sources on the main sequence because all SFGs are above the main sequence and mainly in the starburst regime. Therefore, future ALMA observations with better matched control samples are necessary to expand the parameter spaces.

Zoom In Zoom Out Reset image size

Figure 5 shows the histograms of HST/F814W, and ALMA/1 mm sizes. As shown in Table 2, the mean (median) optical sizes of the SFGs, XAGNs, and IRAGNs, are 1.95 ± 0.74 (2.16 ± 0.74), 1.58 ± 0.50 (1.56 ± 0.50), and 0.42 ± 0.42 (0.41 ± 0.42), respectively. And the mean (median) FIR sizes of the SFGs, XAGNs, and IRAGNs, are 1.58 ± 0.55 (1.48 ± 0.39), 1.07 ± 0.52 (0.89 ± 0.56), 0.75 ± 0.39 (0.68 ± 0.45), respectively. Considering the uncertainties of the sizes and the different measurement techniques as described in Section 2, these numbers imply that the rest-frame FIR sizes of all galaxies seem to be similar or more compact than their rest-frame optical sizes. In particular, the FIR size (0.90 ± 0.21 kpc) of our stacked sources is smaller than the optical one (1.58 ± 0.01). Though there are few IRAGNs, most of them show both very compact optical and FIR sizes. The rest-frame optical and FIR sizes of XAGNs are also relatively small or similar as shown in Figure 5.

Zoom In Zoom Out Reset image size

There are only two objects in the CANDELS field as shown in Table 3, and one XAGN is at z ∼ 2. The HST/F160W measurements are better tracers for optical sizes than the HST/F814W ones at z ∼ 2. The current data shows that the two sizes in HST/F160W are comparable but larger than most sizes in HST/F814W. HST/F160W traces rest-frame the optical size and HST/F814W traces the UV size for the same object, so it is difficult to make a direct comparison. We need size measurements using observations at the same (rest-frame) wavelength for larger numbers of objects to draw a statistical conclusion at z ∼ 2.

Table 3.  Properties of 27 Detected (Upper) and 9 Undetected (Lower) ALMA Band 6/7 High-resolution (01–04) Sources

Project Code	Source Name	Band	Beam Size	${\mathrm{SN}}_{\mathrm{peak}}$	${{\rm{S}}}_{\mathrm{peak}}$	${{\rm{S}}}_{\mathrm{int}}$	${\mathrm{ID}}_{\mathrm{COSMOS}}$	R.A.	Decl.	z	${{\rm{M}}}_{* }$	SFR	${{\rm{R}}}_{\mathrm{ACS}({\rm{g}})}$	${{\rm{R}}}_{\mathrm{WFC}3({\rm{g}})}$	${{\rm{R}}}_{\mathrm{ALMA}({\rm{g}})}$	${{\rm{R}}}_{\mathrm{ACS}({\rm{c}})}$	${{\rm{R}}}_{\mathrm{ALMA}({\rm{c}})}$	${{\rm{R}}}_{\mathrm{ALMA}({\rm{d}})}$	${{\rm{R}}}_{\mathrm{ALMA}(\mathrm{UV})}$	AGN
			$(^{\prime\prime} \times ^{\prime\prime} )$		(mJy/beam)	(mJy)	(deg)	(deg)		(M⊙)	(M⊙ yr−1)	(kpc)	(kpc)	(kpc)	(kpc)	(kpc)	(kpc)	(kpc)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)	(20)	(21)
2013.1.00034.S	lowz_cell8_6496	7	0.33 × 0.24	3.54	0.39 ± 0.11	2.17 ± 0.72	239908	149.97412	1.64352	1.04	11.2 ± 0.1	2.2 ± 0.0	1.4 ± 0.2	⋯	⋯	0.6 ± 0.0	2.7 ± 0.5	2.4 ± 0.8	2.5 ± 0.8	XAGN
2016.1.01604.S	317108	7	0.39 × 0.39	14.61	0.71 ± 0.05	1.64 ± 0.15	317108	150.15256	1.76921	1.56	11.2 ± 0.1	2.3 ± 0.1	⋯	⋯	1.9 ± 0.1	3.6 ± 0.1	2.5 ± 0.1	1.8 ± 0.2	1.4 ± 0.2	⋯
2016.1.01604.S	360903	7	0.34 × 0.29	5.67	0.30 ± 0.05	0.85 ± 0.20	360903	149.66491	1.83649	1.35	11.5 ± 0.1	1.9 ± 0.1	1.2 ± 0.6	⋯	⋯	3.6 ± 0.0	2.2 ± 0.3	1.7 ± 0.4	2.7 ± 0.4	⋯
2016.1.01604.S	405986	7	0.33 × 0.29	8.04	0.36 ± 0.04	2.03 ± 0.29	405986	149.79795	1.90880	1.35	11.3 ± 0.1	2.0 ± 0.1	0.6 ± 0.5	⋯	3.0 ± 0.1	2.4 ± 0.0	3.2 ± 0.3	2.8 ± 0.3	3.2 ± 0.3	⋯
2016.1.01604.S	495291	7	0.39 × 0.39	26.43	1.07 ± 0.04	1.75 ± 0.10	495291	150.62793	2.05289	1.30	11.1 ± 0.1	2.2 ± 0.1	3.4 ± 0.7	⋯	⋯	3.1 ± 0.1	2.1 ± 0.1	1.3 ± 0.1	1.1 ± 0.1	⋯
2016.1.01604.S	500431	7	0.39 × 0.39	13.05	0.53 ± 0.04	1.43 ± 0.15	500431	150.13314	2.06073	1.39	11.0 ± 0.1	1.9 ± 0.1	⋯	⋯	2.5 ± 0.9	3.5 ± 0.1	2.7 ± 0.2	1.8 ± 0.3	1.8 ± 0.3	⋯
2013.1.00034.S	lowz_cell11_120263	7	0.40 × 0.24	9.55	1.73 ± 0.18	2.92 ± 0.45	504172	149.83252	2.06602	1.14	11.4 ± 0.0	2.4 ± 0.0	2.2 ± 0.6	⋯	⋯	2.8 ± 0.0	1.6 ± 0.1	0.8 ± 0.3	0.7 ± 0.3	⋯
2016.1.00735.S	CXID-36	7	0.11 × 0.10	15.19	0.66 ± 0.04	2.69 ± 0.22	553928	150.15841	2.13958	1.83	11.8 ± 0.7	3.0 ± 0.0	0.0 ± 8.3	⋯	⋯	0.4 ± 0.0	0.9 ± 0.0	0.8 ± 0.1	0.9 ± 0.1	XAGN
2016.1.01426.S	PACS_819	6	0.13 × 0.09	13.47	0.14 ± 0.01	0.56 ± 0.05	628146	149.98145	2.25320	1.40	10.8 ± 0.0	2.6 ± 0.0	1.0 ± 0.4	⋯	⋯	7.0 ± 0.1	0.9 ± 0.0	0.8 ± 0.1	0.4 ± 0.1	⋯
2016.1.01604.S	659927	7	0.34 × 0.29	8.05	0.41 ± 0.05	1.17 ± 0.19	659927	150.34101	2.30099	1.44	11.4 ± 0.1	2.0 ± 0.1	2.6 ± 0.6	⋯	⋯	3.8 ± 0.0	2.3 ± 0.2	1.6 ± 0.3	2.0 ± 0.3	⋯
2016.1.01604.S	766157	7	0.33 × 0.29	13.39	0.84 ± 0.06	1.51 ± 0.16	766157	150.00352	2.46146	1.40	11.3 ± 0.1	2.2 ± 0.0	1.1 ± 101.3	⋯	⋯	2.5 ± 0.1	1.8 ± 0.1	1.2 ± 0.2	1.2 ± 0.2	⋯
2013.1.00034.S	midz_cell8_254150	7	0.33 × 0.30	15.95	2.37 ± 0.15	3.09 ± 0.31	794848	150.09342	2.50734	2.20	10.8 ± 0.1	0.8 ± 0.2	1.9 ± 0.7	3.1 ± 0.1	⋯	2.0 ± 0.1	1.5 ± 0.1	0.7 ± 0.3	0.6 ± 0.3	XAGN
2016.1.00778.S	J100054.83 + 023126.2	6	0.45 × 0.32	9.68	0.33 ± 0.03	0.48 ± 0.08	806579	150.23118	2.52466	1.95	10.9 ± 0.1	1.6 ± 0.2	22.3 ± 22.4	⋯	⋯	⋯	1.9 ± 0.1	0.6 ± 0.7	0.6 ± 3.3	⋯
2013.1.00034.S	lowz_cell3_258881	7	0.52 × 0.36	7.44	0.97 ± 0.13	1.64 ± 0.33	811432	150.04247	2.52656	1.23	11.2 ± 0.1	2.1 ± 0.0	1.8 ± 1.1	⋯	2.3 ± 1.0	4.5 ± 0.1	2.4 ± 0.2	1.2 ± 0.7	1.5 ± 0.5	⋯
2013.1.00034.S	midz_cell11_262500	7	0.37 × 0.34	26.26	3.72 ± 0.14	6.70 ± 0.37	815012	150.60329	2.53654	2.19	11.3 ± 0.1	3.2 ± 0.0	1.1 ± 2.0	⋯	⋯	0.8 ± 0.0	2.0 ± 0.1	1.3 ± 0.1	1.2 ± 0.1	⋯
2016.1.01604.S	816106	7	0.39 × 0.39	19.19	0.86 ± 0.05	1.81 ± 0.13	816106	150.08329	2.53609	1.36	11.3 ± 0.1	2.3 ± 0.1	2.3 ± 0.7	3.5 ± 0.0	⋯	3.2 ± 0.0	2.4 ± 0.1	1.7 ± 0.1	1.5 ± 0.1	⋯
2013.1.00034.S	lowz_cell11_268592	7	0.53 × 0.36	3.45	0.36 ± 0.10	2.24 ± 0.75	831023	150.41785	2.55852	1.21	11.0 ± 0.1	2.4 ± 0.0	2.5 ± 0.3	⋯	⋯	2.7 ± 0.0	4.5 ± 0.9	4.1 ± 1.2	3.4 ± 1.4	⋯
2015.1.00026.S	SHIZELS-14	6	0.25 × 0.21	12.96	0.24 ± 0.02	2.33 ± 0.20	831167	150.21496	2.55952	2.08	11.2 ± 0.1	2.7 ± 0.0	2.0 ± 0.8	⋯	⋯	2.3 ± 0.0	3.0 ± 0.2	2.7 ± 0.2	1.6 ± 0.2	⋯
2016.1.00478.S	AzTECC2a	7	0.20 × 0.17	7.72	1.26 ± 0.16	1.78 ± 0.36	842595	149.99797	2.57823	2.42	10.8 ± 0.0	3.0 ± 0.0	0.7 ± 175.9	⋯	⋯	3.6 ± 0.1	0.9 ± 0.1	0.5 ± 0.2	0.7 ± 0.2	XAGN
2013.1.00034.S	lowz_cell11_286398	7	0.54 × 0.37	6.84	0.78 ± 0.11	4.03 ± 0.70	872762	150.12960	2.62144	1.41	11.4 ± 0.1	2.3 ± 0.1	⋯	⋯	⋯	8.2 ± 0.2	4.3 ± 0.4	3.6 ± 0.6	2.4 ± 0.6	⋯
2016.1.01604.S	888769	7	0.39 × 0.39	23.00	1.02 ± 0.04	1.62 ± 0.11	888769	150.10745	2.64591	1.54	10.9 ± 0.2	2.2 ± 0.1	⋯	⋯	2.0 ± 1.1	1.5 ± 0.0	2.1 ± 0.1	1.3 ± 0.1	1.2 ± 0.1	⋯
2016.1.00478.S	AzTECC22a	7	0.19 × 0.18	19.99	3.56 ± 0.18	6.41 ± 0.47	902320	150.03729	2.66960	1.85	11.0 ± 0.1	3.0 ± 0.0	2.3 ± 0.6	⋯	⋯	1.9 ± 0.0	1.0 ± 0.0	0.7 ± 0.1	0.7 ± 0.1	⋯
2016.1.01604.S	917430	7	0.34 × 0.29	13.33	0.70 ± 0.05	0.86 ± 0.11	917430	149.91774	2.69211	1.60	11.1 ± 0.1	2.3 ± 0.1	0.8 ± 0.3	⋯	⋯	0.6 ± 0.0	1.5 ± 0.1	0.6 ± 0.2	0.7 ± 0.2	IRAGN
2013.1.00034.S	lowz_cell10_324567	7	0.56 × 0.38	5.53	0.73 ± 0.13	1.45 ± 0.38	960580	149.91909	2.76053	1.13	10.7 ± 0.1	2.1 ± 0.0	2.3 ± 0.3	⋯	⋯	3.0 ± 0.0	2.7 ± 0.4	1.9 ± 0.7	1.8 ± 0.8	⋯
2013.1.00034.S	lowz_cell10_328266	7	0.57 × 0.38	5.81	0.80 ± 0.14	1.61 ± 0.40	969105	150.53066	2.77548	1.35	11.2 ± 0.0	2.4 ± 0.0	0.1 ± 0.3	⋯	⋯	0.6 ± 0.0	2.8 ± 0.3	2.0 ± 0.6	1.2 ± 0.6	IRAGN
2016.1.01355.S	972308	7	0.12 × 0.10	5.79	0.52 ± 0.09	1.37 ± 0.31	972308	150.40459	2.78053	0.60	10.6 ± 0.1	1.4 ± 0.0	0.0 ± 0.0	⋯	⋯	0.4 ± 0.0	0.6 ± 0.1	0.5 ± 0.1	0.5 ± 0.1	IRAGN
2013.1.00034.S	midz_cell6_332275	7	0.37 × 0.33	19.46	3.00 ± 0.15	7.19 ± 0.50	980250	150.01611	2.79235	1.76	11.3 ± 0.1	2.6 ± 0.0	⋯	⋯	⋯	0.7 ± 0.0	2.3 ± 0.1	1.7 ± 0.1	1.6 ± 0.1	⋯
2016.1.01355.S	263569	7	0.12 × 0.10	⋯	⋯	⋯	263569	149.95048	1.68282	0.78	10.8 ± 0.1	1.3 ± 0.1	2.4 ± 0.4	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	334665	7	0.12 × 0.10	⋯	⋯	⋯	334665	150.43685	1.79439	0.56	10.9 ± 0.1	1.2 ± 0.1	1.8 ± 0.1	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	335028	7	0.12 × 0.10	⋯	⋯	⋯	335028	149.50325	1.79492	0.65	11.3 ± 0.1	1.8 ± 0.1	1.5 ± 0.1	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	410490	7	0.12 × 0.10	⋯	⋯	⋯	410490	150.02499	1.91477	0.97	11.2 ± 0.1	1.9 ± 0.0	⋯	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	414248	7	0.12 × 0.10	⋯	⋯	⋯	414248	149.84769	1.92058	0.71	10.7 ± 0.1	1.2 ± 0.2	0.0 ± 0.0	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	417645	7	0.12 × 0.10	⋯	⋯	⋯	417645	150.02586	1.92639	0.69	10.7 ± 0.1	1.4 ± 0.1	0.7 ± 0.1	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	456808	7	0.12 × 0.10	⋯	⋯	⋯	456808	149.57039	1.99058	1.00	11.2 ± 0.1	1.1 ± 0.2	1.1 ± 0.2	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	711692	7	0.12 × 0.10	⋯	⋯	⋯	711692	149.99343	2.37723	0.93	10.7 ± 0.1	1.6 ± 0.1	0.9 ± 0.3	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN
2016.1.01355.S	755017	7	0.12 × 0.10	⋯	⋯	⋯	755017	150.76449	2.44222	0.70	10.9 ± 0.1	1.8 ± 0.1	3.4 ± 0.2	⋯	⋯	⋯	⋯	⋯	⋯	IRAGN

Note. (1) ALMA project code. (2) ALMA source name. (3) ALMA bands. (4) ALMA beam size. (5) ALMA S/N. (6) ALMA peak flux from IMFIT. (7) Integrated total flux from IMFIT. (8) COSMOS ID. (9) COSMOS R.A. (10) COSMOS decl. (11) Best redshift in COSMOS. (12) Stellar mass estimation from MAGPHYS. The uncertainties of stellar masses are from the probability density functions, which depend on the model templates. The real uncertainty can be larger than our estimation. (13) SFR estimation from MAGPHYS. (14) Sizes measured from HST/F814W images with GALFIT. (15) Sizes measured from HST/F160W images with GALFIT (16) Sizes measured from HST/F160W images with GALFIT. (17) Sizes measured from HST/F814W images with IMFIT (convolved). (18) Sizes measured from ALMA/ 1mm images with IMFIT (convolved). (19) Sizes measured from ALMA/ 1mm images with IMFIT (deconvolved). (20) Sizes measured from ALMA/ 1mm images with UVMODELFIT. (21) IRAGN, XAGN, or non-AGN galaxies.

Download table as:  ASCIITypeset image

Figure 6 shows the ALMA/1 mm sizes to the infrared luminosity relation. Our galaxies fall more or less on top of the Fujimoto et al. (2017) line, but it is difficult to show a clear relation because of the scatter and sample size. Here we decompose the AGN and star formation components according to the SEDs by MAGPHYS+AGN as described in Chang et al. (2017a). Therefore, we are able to derive the infrared luminosity from the star formation and AGN components, respectively. Because FIR luminosity traces dusts, this decomposition can separate dust emission of AGNs from their host galaxies. Figure 6 does not show an obvious correlation between the FIR size and the IR luminosity from the AGN component. According to our SED decomposition, the AGN contribution in the FIR range is negligible (<1% of the total flux), so the ALMA/ 1mm sizes are dominated by star-forming populations. Therefore, the middle panel in Figure 6 is approximate to the FIR size of hosts versus the infrared luminosity from the star-forming part. It is possible that there are correlations between the FIR sizes of the AGNs and the infrared luminosity from the AGN part. However, deeper and higher resolution data are required to resolve the AGN component at 1mm with ALMA.

Zoom In Zoom Out Reset image size

Figure 7 shows the specific SFR (sSFR = SFR/M_*) to SFR surface density (${{\rm{\Sigma }}}_{\mathrm{SFR}}=\mathrm{SFR}/(\pi {R}_{e}^{2})$, where R_e is the rest-frame FIR size) plots. Because there are more massive and star-forming sources at high redshifts, it is clear to see that high redshift sources having higher SFR will naturally imply higher SFR surface density. It was also mentioned earlier that the sizes were systematically smaller at higher redshift, which would also increase the surface densities even at a fixed SFR or a fixed sSFR. At z ∼ 1, the SFR surface density of our obscured IRAGN hosts are high compared to archived ALMA observations at a given sSFR. At z ∼ 2, the SFR surface density of both XAGN and IRAGN hosts seem to be similar or slightly higher than that of SFGs, but the uncertainties of the SFR surface density are large. In Figure 7, SFR surface densities of our IRAGN hosts are much higher than SFGs on the main sequence at both z ∼ 1. This suggests that the compactness of IRAGN hosts will be significant if we have a large control sample of SFGs. Therefore, more data with high resolution are needed to reach a statistically meaningful conclusion.

Zoom In Zoom Out Reset image size

4.1. Compact Rest-frame FIR and Optical sizes

4.2. FIR versus Optical Sizes in AGN Hosts

4.3. SFR Surface Densities of Obscured IRAGNs

In this paper, we have investigated rest-frame FIR and optical sizes of non-AGN SFGs, infrared-selected AGNs, and X-ray selected AGNs at z < 2.5 by using high-resolution ALMA (010–085) and HST imaging (∼01) in the COSMOS field. Our main findings are as follows:

• 1.  
  The sizes measured from both ALMA/1 mm and HST/F814W images show that obscured AGN host galaxies at z ∼ 1 are more compact than SFGs at a given redshift and stellar mass.
• 2.  
  Most detected rest-frame FIR sizes of the sources are similar or more compact than their rest-frame optical sizes, which is consistent with recent results for ALMA-detected sources. This can be explained by the fact that the dusty starbursts take place in the compact regions, and suggests that the star formation mechanisms in the compact regions of AGN hosts are similar to those taking place in SFGs with no evidence of an AGN at their center.
• 3.  
  The decomposed IRAGN luminosity may imply that AGN activity has little influence on the FIR sizes. However, more high-resolution imaging from future ALMA observations are needed to confirm this result.
• 4.  
  The median SFR surface density of the obscured AGN hosts is higher than those of other ALMA-detected sources at $\mathrm{log}\,{M}_{* }/{M}_{\odot }\sim 11$ and z ∼ 1. This implies that obscured AGN host galaxies may be in a transition phase of compact SFGs, and may evolve into compact quiescent galaxies at later time.
• 5.  
  Future high-resolution ALMA observations would be crucial to investigating these compact AGN host galaxies further.

We thank the referee, as well as J. Freundlich for helpful comments and discussions. Y.Y.C. acknowledges financial support from the Ministry of Science and Technology of Taiwan grant (105-2112-M-001-029-MY3 and 108-2112-M-001-014-), and the Agence Nationale de la Recherche (contract #ANR-12-JS05-0008-01). E.d.C. gratefully acknowledges the Australian Research Council as the recipient of a Future Fellowship (project FT150100079). This paper makes use of the following ALMA data: ADS/JAO.ALMA #2013.1.00034.S, #2015.1.00026.S, #2016.1.00478.S, #2016.1.00735.S, #2016.1.00778.S, #2016.1.01355.S, #2016.1.01426.S, and #2016.1.01604.S. ALMA is a partnership of ESO (representing its member states), NSF (USA), and NINS (Japan), together with NRC (Canada), MOST and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO, and NAOJ. We gratefully acknowledge the contributions of the entire COSMOS collaboration.