The Role of Major Mergers and Nuclear Star Formation in Nearby Obscured Quasars

Our type 2 quasars were originally selected to complement a matching study of low-redshift type 1 quasars selected from the Palomar-Green survey of Schmidt & Green (1983) to examine the evolutionary connection between these two populations, which will be reported in an upcoming paper. The comparison sample of Palomar-Green quasars consists of 87 objects with z < 0.5 (Boroson & Green 1992). To this end, we randomly selected 87 type 2 quasars, matching the Palomar-Green sample in terms of redshift and [O iii] λ5007 luminosity, from the catalog of 887 type 2 quasars published by Reyes et al. (2008).⁶ Type 2 quasars in this catalog were identified from the Sloan Digital Sky Survey (SDSS; York et al. 2000) Data Release 6 spectroscopic database (Adelman-McCarthy et al. 2008). Our selection assumes that type 2 quasars have the same intrinsic AGN luminosity as type 1 quasars for a given L_{[O iii]}, and that L_{[O iii]} is related to the bolometric luminosity of the AGN (Heckman et al. 2005; LaMassa et al. 2009; Dicken et al. 2014).

As the observations were conducted in the "snapshot" mode of HST, only ∼1/3 of the original sample of type 2 quasars was observed successfully, yielding a sample of 29 objects (Table 1). The final sample spans a redshift range of 0.04–0.4, with a median value of z = 0.12, and extinction-corrected [O iii] luminosities from $\mathrm{log}({L}_{[{\rm{O}}{\rm{III}}]}/{L}_{\odot })=8.42$ to 9.88, with a median value of 9.11 (Figure 1). The 29 objects have a similar distribution of redshifts, [O iii] luminosities, and optical magnitudes as the original sample of 87 objects. They also match the distribution of these quantities in the parent catalog of Reyes et al. (2008) at z < 0.5.

Zoom In Zoom Out Reset image size

Table 1.  Observational Information

Object	z	DL	E(B − V)	$\mathrm{log}{L}_{[{\rm{O}}{\rm{III}}]}$	$\mathrm{log}{L}_{[{\rm{O}}{\rm{III}}]}$	Filter	ExpTime	ObsDate
		(Mpc)	(mag)	(L⊙)	(L⊙)	(IWFC3/BWFC3)	(s)	(yy mm dd)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
SDSS J011935.63−102613.1	0.125	586	0.0371	8.43	8.81	F105W/F475W	147/780	2013-06-19
SDSS J074751.56+320052.1	0.280	1438	0.0700	8.95	9.88	F110W/F555W	147/780	2014 Apr 27
SDSS J075329.93+230930.7	0.336	1775	0.0633	8.59	8.92	F110W/F555W	147/780	2014 Feb 13
SDSS J075940.95+505024.0	0.054	243	0.0415	8.83	9.32	F814W/F438W	470/300	2013 Sep 9
SDSS J080252.92+255255.5	0.081	368	0.0329	8.86	9.23	F105W/F438W	147/705	2013 Sep 13
SDSS J080337.32+392633.1	0.065	295	0.0456	8.12	9.25	F105W/F438W	147/705	2014 Jan 12
SDSS J080523.29+281815.8	0.128	604	0.0476	8.62	9.14	F105W/F475W	147/570	2013 Jan 3
SDSS J081100.20+444216.3	0.183	888	0.0418	8.27	9.85	F105W/F475W	147/540	2013 Jan 21
SDSS J084107.06+033441.3	0.274	1404	0.0334	8.77	9.36	F110W/F555W	147/780	2014 Jan 21
SDSS J084344.99+354941.9	0.054	241	0.0360	8.14	8.55	F814W/F438W	260/320	2013 Feb 23
SDSS J090754.07+521127.5	0.085	386	0.0159	8.23	8.72	F105W/F438W	147/720	2013 Mar 12
SDSS J091819.66+235736.4	0.419	2303	0.0443	9.58	9.87	F125W/F555W	147/780	2013 Apr 14
SDSS J093625.36+592452.7	0.095	439	0.0200	8.35	8.63	F105W/F438W	147/780	2013 Apr 28
SDSS J103408.59+600152.2	0.050	225	0.0093	8.85	9.14	F814W/F438W	156/130	2013 Oct 9
SDSS J105208.19+060915.1	0.052	232	0.0310	8.20	8.60	F814W/F438W	325/390	2013 Feb 25
SDSS J110213.01+645924.8	0.077	352	0.0319	8.45	9.54	F105W/F438W	147/660	2013 Aug 3
SDSS J111015.25+584845.9	0.143	678	0.0096	8.88	9.08	F105W/F475W	147/780	2014 Mar 21
SDSS J113710.77+573158.7	0.395	2152	0.0097	9.61	9.87	F125W/F555W	147/780	2014 Mar 21
SDSS J115326.42+580644.5	0.064	290	0.0249	8.48	8.97	F105W/F438W	147/630	2014 Mar 19
SDSS J123804.81+670320.7	0.179	871	0.0190	8.26	8.69	F105W/F475W	147/780	2012 Dec 31
SDSS J125850.77+523913.0	0.055	246	0.0141	8.26	8.42	F814W/F438W	325/390	2013 Jan 11
SDSS J130038.09+545436.8	0.088	403	0.0180	8.94	9.11	F105W/F438W	147/660	2013 Aug 6
SDSS J133542.49+631641.5	0.169	816	0.0187	8.47	9.49	F105W/F475W	147/780	2013 Aug 10
SDSS J140541.21+402632.5	0.080	366	0.0135	8.78	9.17	F105W/F438W	147/660	2014 Aug 07
SDSS J140712.94+585120.4	0.170	823	0.0107	8.27	8.80	F105W/F475W	147/780	2013 May 29
SDSS J144038.09+533015.8	0.038	166	0.0115	8.94	9.30	F814W/F438W	188/132	2013 Nov 24
SDSS J145019.18−010647.4	0.120	559	0.0459	8.42	8.61	F105W/F475W	147/660	2013 Jun 15
SDSS J155829.36+351328.6	0.119	558	0.0246	8.77	8.98	F105W/F475W	147/540	2012 Oct 27
SDSS J162436.40+334406.7	0.122	573	0.0227	8.56	8.82	F105W/F475W	147/660	2013 Sep 7

Note. Column (1): Object name. Column (2): Redshift. Column (3): Luminosity distance. Column (4): Galactic extinction. Column (5): Observed [O iii] luminosity (Reyes et al. 2008). Column (6): Extinction-corrected [O iii] luminosity (Kong & Ho 2018). Column (7): WFC3 filter. Column (8): Exposure time for I_WFC3 and B_WFC3. Column (9): Date of observations.

Download table as:  ASCIITypeset image

The observations were conducted using the WFC3 camera between 2012 November and 2014 July (proposal ID 12903; PI: Luis C. Ho). Each object was observed with a blue filter and a red filter using the UVIS or IR channel. The bandpasses were carefully chosen from the large suite of available WFC3 filters, with two considerations: to match approximately rest-frame B and I, and to avoid strong emission lines. Hereinafter, we denote the bluer filter (F438W, F475W, F555W) as B_WFC3 and the redder filter (F814W, F105W, F110W, F125W) as I_WFC3.

For the UVIS channel, each quasar was observed with three long exposures, using the three-point dithering pattern WFC3-UVIS-DITHER-LINE-3PT. For the IR channel, four long exposures were taken using the four-point dithering pattern WFC3-IR-DITHER-BOX-MIN. We used subarrays to minimize the readout time and buffer size, resulting in a field-of-view (FoV) of 67 × 67 arcsec² and 40 × 40 arcsec² for the IR and UVIS channels, respectively. These FoVs, which correspond to ∼145 and 87 kpc at the median redshift of the sample, are sufficiently wide to cover the outskirts of the host galaxies for detecting extended features and to achieve accurate sky measurement. Total exposure times, which varied between 147 and 780 s, were set to reach a surface brightness limit of μ ≈ 25 mag arcsec⁻², a depth that previous studies had demonstrated can yield robust detections of faint outer structures (e.g., Greene et al. 2008; Kim et al. 2008a; Jiang et al. 2011). Table 1 gives a summary of the observations.

We use AstroDrizzle to combine the dithered images to generate cosmic ray-removed science images. The pixel scale is set to 006 for the IR channel and 003 for the UVIS channel so that it Nyquist samples the point-spread function (PSF), which has a full width at half maximum (FWHM) of ∼013 and 007 for the IR and UVIS channel, respectively. Fortunately, none of the central pixels near the nucleus was saturated. Although AstroDrizzle performs sky subtraction, further adjustments of the sky level were made during the two-dimensional (2D) image fitting process using GALFIT (Peng et al. 2002, 2010; see Section 3.2.1).

A robust model of the PSF is crucial for accurate image decomposition, even for the hosts of type 2 AGNs. Ideally, the PSF can be constructed from bright stars observed simultaneously in the science images, but in general, this is not possible in our program because of the relatively small FoV of our subarray images. While synthetic TinyTim (Krist et al. 2011) PSFs are commonly used as a substitute (e.g., Kim et al. 2017), our experience (Huang et al. 2019) indicates that empirical PSFs generated from stacked WFC3 images of multiple bright, unsaturated, isolated stars observed with the same filter, in the same dither pattern, but at different times perform significantly better than synthetic PSFs. Hence, our analysis uses a library of empirical PSFs of high signal-to-noise ratio (S/N) created from stacking individual stars. The total number of stars used to generate the stacked PSF differs from filter to filter, with the average being a few tens.

The high resolution and sensitivity of the WFC3 images, coupled with the low redshifts of our sample, allow us to perform quite reliable visual classifications of the galaxy morphologies. We distinguish five broad types: merging/disturbed, unbarred spirals, barred spirals, lenticulars, and ellipticals. In this work, we regard spirals and lenticulars as disk galaxies, and we consider spirals as late-type.

We summarize the classifications of the 29 objects as follows (Figure 2): 10 can be considered merging/disturbed because they exhibit obvious signs of interactions, distorted features, or otherwise reside in host galaxies in close pairs; three are found in unbarred and eight in barred spirals; five are hosted by lenticulars; and the remaining three are in ellipticals. The morphological type of each quasar can be found in Table 2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 2.  Best-fit Parameters of Host Galaxies

	IWFC3									BWFC3
Object	Filter	nbulge	${R}_{e,\mathrm{bulge}}$	Bulge	Disk	Bar	Nucleus	Total	B/T	Filter	Bulge	Disk	Bar	Nucleus	Total	Morphology
			('')	(mag)	(mag)	(mag)	(mag)	(mag)			(mag)	(mag)	(mag)	(mag)	(mag)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)
J011935	F105W	${1.85}_{-0.20}^{+0.19}$	${0.78}_{-0.08}^{+0.08}$	${17.03}_{-0.09}^{+0.09}$	${16.58}_{-0.01}^{+0.01}$	⋯	${19.52}_{-0.03}^{+0.09}$	15.97	0.27	F475W	${19.94}_{-0.11}^{+0.12}$	${18.85}_{-0.02}^{+0.02}$	⋯	${21.42}_{-0.09}^{+0.13}$	18.51	spiral (unbarred)
J074751	F110W	${0.53}_{-0.12}^{+0.11}$	${0.09}_{-0.05}^{+0.04}$	${18.83}_{-0.10}^{+0.10}$	${16.98}_{-0.01}^{+0.03}$	${18.70}_{-0.02}^{+0.03}$	⋯	16.62	0.13	F555W	${20.92}_{-0.11}^{+0.11}$	${19.18}_{-0.01}^{+0.02}$	${21.65}_{-0.02}^{+0.03}$	⋯	18.89	spiral (barred)
J075329	F110W	${2.25}_{-0.47}^{+0.46}$	${0.80}_{-0.16}^{+0.16}$	${18.20}_{-0.19}^{+0.19}$	${19.66}_{-0.03}^{+0.01}$	⋯	${20.05}_{-0.02}^{+0.02}$	17.95	0.79	F555W	${21.05}_{-0.21}^{+0.22}$	${22.31}_{-0.02}^{+0.03}$	⋯	${23.98}_{-0.01}^{+0.14}$	20.76	merging
J075940b	F814W	${1.04}_{-0.20}^{+0.12}$	${0.12}_{-0.02}^{+0.01}$	${16.60}_{-0.09}^{+0.09}$	${15.48}_{-0.03}^{+0.03}$	${16.38}_{-0.01}^{+0.01}$	⋯	14.85	0.20	F438W	${18.17}_{-0.09}^{+0.12}$	${17.62}_{-0.07}^{+0.06}$	${18.84}_{-0.04}^{+0.05}$	⋯	16.91	lenticular
J080252	F105W	${4.53}_{-0.92}^{+0.91}$	${1.65}_{-0.33}^{+0.33}$	${14.16}_{-0.14}^{+0.16}$	${14.76}_{-0.00}^{+0.02}$	⋯	⋯	13.67	0.63	F438W	${17.25}_{-0.17}^{+0.19}$	${17.69}_{-0.00}^{+0.01}$	⋯	⋯	16.70	merging
J080337	F105W	${4.57}_{-0.46}^{+0.46}$	${4.82}_{-0.52}^{+0.52}$	${13.68}_{-0.07}^{+0.07}$	${15.16}_{-0.04}^{+0.04}$	⋯	⋯	13.43	0.80	F438W	${18.68}_{-0.12}^{+0.15}$	${17.19}_{-0.04}^{+0.03}$	⋯	⋯	16.94	lenticular
J080523a	F105W	${2.22}_{-0.49}^{+0.45}$	${0.87}_{-0.18}^{+0.18}$	${16.87}_{-0.17}^{+0.18}$	${15.40}_{-0.00}^{+0.01}$	⋯	${18.05}_{-0.07}^{+0.05}$	15.15	0.21	F475W	${20.23}_{-0.21}^{+0.22}$	${17.91}_{-0.02}^{+0.04}$	⋯	${19.94}_{-0.09}^{+0.01}$	17.79	merging
J081100	F105W	${4.03}_{-0.56}^{+0.48}$	${0.41}_{-0.06}^{+0.05}$	${16.97}_{-0.10}^{+0.12}$	${16.07}_{-0.01}^{+0.02}$	⋯	⋯	15.61	0.18	F475W	${20.05}_{-0.10}^{+0.11}$	${18.30}_{-0.03}^{+0.02}$	⋯	⋯	18.10	spiral (unbarred)
J084107	F110W	${4.08}_{-0.87}^{+0.85}$	${1.46}_{-0.30}^{+0.30}$	${17.07}_{-0.17}^{+0.17}$	⋯	⋯	${19.65}_{-0.06}^{+0.19}$	17.07	1.00	F555W	${19.39}_{-0.19}^{+0.19}$	⋯	⋯	${23.30}_{-0.08}^{+0.02}$	19.39	merging
J084344	F814W	${3.98}_{-1.23}^{+0.89}$	${4.18}_{-1.99}^{+1.57}$	${14.67}_{-0.42}^{+0.25}$	${15.12}_{-0.24}^{+0.10}$	⋯	⋯	14.12	0.60	F438W	${16.97}_{-0.17}^{+0.18}$	${17.66}_{-0.08}^{+0.08}$	⋯	⋯	16.51	disturbed
J090754	F105W	${1.89}_{-0.21}^{+0.21}$	${0.24}_{-0.02}^{+0.02}$	${16.42}_{-0.08}^{+0.09}$	${15.49}_{-0.01}^{+0.01}$	${16.22}_{-0.01}^{+0.01}$	${19.46}_{-0.20}^{+0.48}$	14.77	0.22	F438W	${19.39}_{-0.11}^{+0.10}$	${18.44}_{-0.11}^{+0.10}$	${19.56}_{-0.05}^{+0.06}$	${23.24}_{-0.05}^{+0.12}$	17.82	spiral (barred)
J091819	F125W	${1.79}_{-0.25}^{+0.14}$	${1.10}_{-0.07}^{+0.08}$	${18.75}_{-0.06}^{+0.08}$	⋯	⋯	${19.26}_{-0.06}^{+0.14}$	18.75	1.00	F555W	${20.95}_{-0.06}^{+0.06}$	⋯	⋯	${21.01}_{-0.01}^{+0.15}$	20.95	elliptical
J093625b	F105W	${2.92}_{-0.31}^{+0.29}$	${0.17}_{-0.02}^{+0.02}$	${16.81}_{-0.09}^{+0.09}$	${16.01}_{-0.01}^{+0.01}$	${17.56}_{-0.04}^{+0.02}$	⋯	15.42	0.28	F438W	${19.77}_{-0.10}^{+0.10}$	${19.39}_{-0.11}^{+0.10}$	${17.90}_{-0.02}^{+0.01}$	⋯	17.51	lenticular
J103408	F814W	${3.42}_{-1.55}^{+0.79}$	${3.99}_{-1.56}^{+1.10}$	${14.45}_{-1.52}^{+0.35}$	${14.82}_{-0.64}^{+0.18}$	⋯	⋯	13.87	0.59	F438W	${16.31}_{-0.16}^{+0.17}$	${17.35}_{-0.11}^{+0.10}$	⋯	⋯	15.96	disturbed
J105208	F814W	${2.08}_{-0.22}^{+0.22}$	${0.49}_{-0.05}^{+0.05}$	${16.95}_{-0.09}^{+0.09}$	${15.39}_{-0.03}^{+0.03}$	${16.37}_{-0.05}^{+0.03}$	${19.59}_{-0.07}^{+0.05}$	14.85	0.14	F438W	${19.48}_{-0.10}^{+0.10}$	${17.38}_{-0.07}^{+0.06}$	${18.86}_{-0.04}^{+0.07}$	${20.49}_{-0.05}^{+0.07}$	17.01	spiral (barred)
J110213a	F105W	${2.06}_{-0.47}^{+0.41}$	${1.58}_{-0.32}^{+0.32}$	${15.11}_{-0.16}^{+0.16}$	${15.01}_{-0.00}^{+0.01}$	⋯	${16.57}_{-0.03}^{+0.16}$	14.30	0.48	F438W	${19.08}_{-0.19}^{+0.19}$	${17.76}_{-0.01}^{+0.00}$	⋯	${20.44}_{-0.09}^{+0.08}$	17.48	merging
J111015	F105W	${4.56}_{-0.45}^{+0.24}$	${0.13}_{-0.01}^{+0.01}$	${17.94}_{-0.04}^{+0.05}$	⋯	⋯	⋯	17.94	1.00	F475W	${19.59}_{-0.07}^{+0.09}$	⋯	⋯	⋯	19.59	elliptical
J113710	F125W	${5.97}_{-0.40}^{+0.38}$	${0.97}_{-0.08}^{+0.08}$	${17.16}_{-0.05}^{+0.05}$	⋯	⋯	⋯	17.16	1.00	F555W	${19.81}_{-0.06}^{+0.06}$	⋯	⋯	⋯	19.81	elliptical
J115326	F105W	${1.75}_{-0.27}^{+0.19}$	${0.19}_{-0.02}^{+0.02}$	${16.66}_{-0.08}^{+0.11}$	${15.34}_{-0.01}^{+0.01}$	${16.54}_{-0.01}^{+0.01}$	${18.04}_{-0.02}^{+0.26}$	14.81	0.18	F438W	${19.31}_{-0.10}^{+0.10}$	${17.95}_{-0.07}^{+0.07}$	${19.81}_{-0.07}^{+0.09}$	${20.22}_{-0.04}^{+0.01}$	17.54	spiral (barred)
J123804	F105W	${1.31}_{-0.13}^{+0.13}$	${0.20}_{-0.02}^{+0.02}$	${19.12}_{-0.10}^{+0.10}$	${16.93}_{-0.00}^{+0.00}$	${16.91}_{-0.00}^{+0.00}$	⋯	16.10	0.06	F475W	${22.17}_{-0.12}^{+0.12}$	${20.13}_{-0.00}^{+0.02}$	${18.83}_{-0.04}^{+0.02}$	⋯	18.51	spiral (barred)
J125850	F814W	${1.62}_{-0.33}^{+0.16}$	${0.61}_{-0.06}^{+0.06}$	${16.57}_{-0.09}^{+0.11}$	${14.88}_{-0.05}^{+0.05}$	${16.43}_{-0.02}^{+0.03}$	⋯	14.48	0.15	F438W	${18.85}_{-0.10}^{+0.12}$	${17.24}_{-0.15}^{+0.13}$	${19.60}_{-0.14}^{+0.16}$	⋯	16.92	spiral (barred)
J130038b	F105W	${3.82}_{-0.41}^{+0.42}$	${0.71}_{-0.08}^{+0.09}$	${15.90}_{-0.10}^{+0.09}$	${15.79}_{-0.02}^{+0.01}$	${16.86}_{-0.09}^{+0.09}$	⋯	14.90	0.40	F438W	${18.43}_{-0.10}^{+0.10}$	${19.85}_{-1.85}^{+0.65}$	${19.16}_{-0.09}^{+0.09}$	⋯	17.80	lenticular
J133542	F105W	${3.01}_{-0.61}^{+0.61}$	${1.69}_{-0.34}^{+0.34}$	${16.06}_{-0.16}^{+0.16}$	⋯	⋯	${18.58}_{-0.01}^{+0.05}$	16.06	1.00	F475W	${18.33}_{-0.18}^{+0.19}$	⋯	⋯	${21.93}_{-0.03}^{+0.17}$	18.33	disturbed
J140541	F105W	${1.69}_{-0.20}^{+0.17}$	${0.50}_{-0.05}^{+0.05}$	${15.83}_{-0.08}^{+0.09}$	${16.16}_{-0.01}^{+0.02}$	⋯	${17.60}_{-0.01}^{+0.07}$	15.19	0.45	F438W	${18.16}_{-0.10}^{+0.10}$	${18.65}_{-0.10}^{+0.09}$	⋯	${19.70}_{-0.03}^{+0.11}$	17.63	spiral (unbarred)
J140712	F105W	${1.86}_{-0.19}^{+0.19}$	${0.19}_{-0.02}^{+0.02}$	${17.28}_{-0.09}^{+0.09}$	${16.80}_{-0.01}^{+0.01}$	${17.44}_{-0.03}^{+0.03}$	⋯	15.94	0.29	F475W	${19.66}_{-0.10}^{+0.13}$	${19.00}_{-0.04}^{+0.04}$	${20.39}_{-0.04}^{+0.04}$	⋯	18.35	spiral (barred)
J144038	F814W	${2.04}_{-0.45}^{+0.42}$	${0.19}_{-0.05}^{+0.04}$	${16.35}_{-0.16}^{+0.17}$	${14.34}_{-0.03}^{+0.03}$	⋯	${17.96}_{-0.04}^{+0.16}$	14.18	0.14	F438W	${17.54}_{-0.18}^{+0.19}$	${16.09}_{-0.04}^{+0.04}$	⋯	${18.64}_{-0.08}^{+0.03}$	15.84	merging
J145019	F105W	${4.22}_{-0.50}^{+0.54}$	${1.81}_{-0.28}^{+0.39}$	${15.38}_{-0.12}^{+0.14}$	${16.88}_{-0.20}^{+0.36}$	⋯	⋯	15.14	0.80	F475W	${18.51}_{-0.10}^{+0.12}$	${18.89}_{-0.06}^{+0.05}$	⋯	⋯	17.93	lenticular
J155829	F105W	${0.80}_{-0.09}^{+0.08}$	${0.31}_{-0.03}^{+0.03}$	${18.10}_{-0.09}^{+0.09}$	${16.25}_{-0.01}^{+0.01}$	${17.11}_{-0.03}^{+0.03}$	${19.48}_{-0.00}^{+0.07}$	15.72	0.11	F475W	${19.99}_{-0.10}^{+0.10}$	${18.30}_{-0.05}^{+0.05}$	${19.82}_{-0.05}^{+0.05}$	${21.67}_{-0.03}^{+0.13}$	17.89	spiral (barred)
J162436	F105W	${1.93}_{-0.41}^{+0.39}$	${0.79}_{-0.16}^{+0.16}$	${16.84}_{-0.18}^{+0.18}$	${15.88}_{-0.00}^{+0.01}$	⋯	${19.14}_{-0.02}^{+0.08}$	15.51	0.29	F475W	${19.27}_{-0.19}^{+0.20}$	${18.59}_{-0.04}^{+0.04}$	⋯	${21.31}_{-0.05}^{+0.12}$	18.13	merging

Notes. Column (1): Galaxy name. Column (2): Filter used, corresponding to I_WFC3. Columns (3)−(8): Best-fit GALFIT parameters. Column (9): Total magnitude of the host, excluding the nucleus, if present. Column (10): B/T ratio. Column (11): Filter used, corresponding to B_WFC3. Columns (12)−(15): Best-fit GALFIT parameters. Column (16): Total magnitude of the host, excluding the nucleus, if present. Column (17): Morphological classification.

^aThe magnitude of the disk in Columns (6) and (13) is the sum of the magnitudes of all the components except the bulge and nucleus component. ^bThis galaxy has an inner lens; the magnitude given in Columns (7) and (14) refer to the lens instead of the bar.

Download table as:  ASCIITypeset image

Both the merging/disturbed system SDSS J162436.40+334406.7 and the lenticular galaxy SDSS J093625.36+592452.7 have an apparent close companion. Based on available redshift measurements, the companion of SDSS J162436.40+334406.7 is genuinely associated with it, thus indeed constituting a merging system. However, the apparent companion of SDSS J093625.36+592452.7 is a foreground galaxy at z = 0.04. While the WFC3 images of SDSS J144038.09+533015.8 suggest that the host galaxy is an isolated spiral, a larger FoV SDSS image (see inset panel in Figure 2) reveals a long stellar bridge connecting the quasar to a disturbed companion, which prompted us to classify the host as merging/disturbed. We inspected large-FoV SDSS images for all the other objects in the sample and found no other examples of potential companions.

Surprisingly, the majority of the sample (55%) exhibit unambiguous large-scale disks, with a significant fraction (38%) hosting clear late-type morphologies in the form of spiral arms and bars. Only approximately one-third of the host galaxies reside in close pairs or show obvious signatures of ongoing or recent interactions. If the ellipticals can be considered merger products, then in total, ∼45% of the sample is or has been associated with mergers of one type or another. Section 4.2 discusses the implications of these findings in relation to quasar triggering mechanisms.

Figure 3 shows the redshift distribution of the host galaxies with different morphological types. The galaxies with disks and morphologically disturbed features generally concentrate toward lower redshifts (z ≈ 0.1) than those classified as ellipticals (median z ≈ 0.4). Have extended features of low surface brightness been missed in these more distant objects?

Zoom In Zoom Out Reset image size

To test whether surface brightness dimming is the main cause of the classification of the elliptical galaxies, we use the FERENGI code (Barden et al. 2008) to generate mock, redshifted images of galaxies using the actual observed images of lower redshift objects. To match the luminosity of quasars with elliptical hosts ($\mathrm{log}{L}_{[{\rm{O}}{\rm{III}}]}/{L}_{\odot }=8.88,9.58,9.61$), we choose nearby counterparts with similar [O iii] luminosities covering the range of extended morphologies (merging/disturbed, barred and unbarred spirals, lenticulars). The code takes into account the cosmological corrections for size, surface brightness, bandpass shifting, and k-correction. FERENGI treats the cosmological evolution of the stellar population, crudely parameterizing the luminosity evolution as dM/dz = −1 (Ilbert et al. 2005). We assume that the mock high-z quasars are located at z = 0.143 and observed with the filter pair F475W/F105W, or at z = 0.4 and observed with the filter pair F555W/F125W.

In the set of simulated IR images (upper panel in Figure 4), many of the original morphological details are lost, and most of the galaxies would be incorrectly classified as ellipticals or lenticulars when viewed in F105W at z = 0.143 or in F125W at z = 0.4. However, the simulated UVIS F475W and F555W images do a much better job in retaining the original structural information (bottom panel in Figure 4). Therefore, so long as images in both filters are available, the morphological classification is unlikely to be biased in our study. We conclude that the morphological classifications of the three elliptical hosts in our sample should be secure.

Zoom In Zoom Out Reset image size

We analyze the images using GALFIT V3.0 (Peng et al. 2010), a nonlinear least-squares fitting code that uses Levenberg-Marquardt minimization to decompose the major structural components of galaxies. We allow GALFIT to generate its own σ (weight) image. The sky value is fixed to a constant determined from the average background value of five source-free 150 × 150 pixel² regions, and the uncertainty of the sky is the standard deviation of the five measurements. The sky measurement is measured separately for each image of each filter. To minimize contamination from nearby sources, we masked all objects beyond 1.5 times the Kron radius⁷ from the target quasar. Additionally, objects that are more than 2.5 mag fainter than the target quasar are masked out regardless of their position because they will hardly affect the fit for the primary target. Unmasked close companions are simultaneously fit with the target galaxy. Prominent dust lanes and regions need to be masked in some objects (e.g., SDSS J080337.32+392633.1, J145019.18−010647.4, and J084344.99+354941.9).

3.2.1. Best-fit Models

We fit bulges with the Sérsic (1968) profile. We set an upper limit of n = 8 for the Sérsic index, which is close to the largest values seen in the most luminous ellipticals (e.g., Kormendy et al. 2009). Moreover, values larger than n = 8 are often associated with poor model fits (Barden et al. 2012). We adopt an exponential profile (n = 1) for the disk component. Spiral arms, when clearly visible, are modeled by coordinate rotation and bending modes provided by GALFIT V3.0 (Peng et al. 2010; additional examples can be found in Gao & Ho 2017 and Gao et al. 2019). Three lenticular galaxies (SDSS J075940.95+505024.0, J093625.36+592452.7, and J130038.09+545436.8) exhibit an inner lens. Gao & Ho (2017, see also Gao et al. 2018) demonstrate that neglecting inner lenses will bias the derived bulge parameters significantly. Therefore, we model the inner lens as an independent component with either a Sérsic or an exponential profile, depending on which gives the better fit.

The bar, when present, needs to be properly included to avoid incurring large errors on the derived properties of the bulge (Laurikainen et al. 2004, 2005; Gadotti 2008; Gao & Ho 2017). Following common practice (e.g., de Jong 1996; Freeman 1966), we adopt a fixed Sérsic n = 0.5 profile for the bar component. Figure 5 demonstrates the effect of the bar component for one of the objects in the sample. Without a bar component (middle panel), the size, brightness, and ellipticity of the bulge appear to be substantially overestimated due to the influence of the bar. The poor match of the ellipticity profile as well as the large residuals at r ≈ 2''−4'' further attest to the inadequacy of the model without a bar.

Zoom In Zoom Out Reset image size

For host galaxies classified as merging/disturbed, especially for those with substantial disturbance, the bulge component cannot always be distinguished clearly from the other irregular components. We take as the bulge the prominent central component, which is usually well fit with a single Sérsic profile (with n ≤ 8). We use Fourier modes (Peng et al. 2010) to model asymmetric structures such as lopsided features and tidal tails.

Although the active nucleus should be deeply obscured in type 2 quasars, a fraction of its light can still scatter out (Antonucci & Miller 1985) and thereby modify the innermost light profile of the galaxy. A significant fraction of our sample requires an additional compact, nuclear component—modeled as an unresolved PSF component—to achieve a satisfactory fit for the bulge. Absent the nuclear component, the Sérsic index of the bulge can reach unrealistically high values that are inconsistent with the values expected for the morphological types of the host galaxies. An example is illustrated in the right panel of Figure 5. The three-component (bulge, bar, and disk) model yields a bulge Sérsic index of n = 4.63, which would be unprecedented for such an obviously late-type galaxy (e.g., Balcells et al. 2003; Vika et al. 2015). The origin of the large Sérsic index is clear: the very central region contains a sharp spike, which has the effect of mimicking a large n. After including an additional nucleus component (left panel of Figure 5), the bulge Sérsic index drops to a much more reasonable value of n = 2.08. This suggests that accounting for a nuclear component, probably due to scattered light, is essential to deriving accurate photometric properties of the bulge. A nuclear component seems to be required in 14 objects, of which six are spirals, two are ellipticals, and six are merging/disturbed hosts. The nucleus typically contributes ≲10% of the total brightness of the galaxy (average of 8% in I_WFC3 and 9% in B_WFC3). Note that, with absolute magnitudes of −18 to −21, these central components are unlikely to be nuclear star clusters, which have typical absolute magnitudes of M_I ≈ −10 to −14 (Böker et al. 2002).

The I_WFC3-band images are significantly deeper than the B_WFC3-band images. The redder bandpass is also intrinsically more sensitive to the dominant, older stellar component of the host, and, of course, is less affected by dust extinction. We first determine the best-fit model using the I_WFC3-band image. Then we fix all structural parameters of the sub-components (e.g., R_e, position angle, ellipticity, central position) to solve only for their brightnesses in the B_WFC3 band. The mask and sky level of the B_WFC3-band image are determined in the same manner as the I_WFC3-band image.

The best-fit models from the I_WFC3-band images for the sample are given in Appendix, and the final parameters are summarized in Table 2. The fits are generally good, with reduced χ² ≈ 1. Quasars that are classified as barred and unbarred spirals tend to have less centrally concentrated (n ≲ 2) and smaller (R_e ≲ 06) bulges, with B/T generally less than 0.2. In contrast, the ellipticals and bulges of lenticular and merging/disturbed hosts have much more concentrated (n > 2), larger (R_e > 06), and more dominant (B/T > 0.2) spheroids. We will discuss in detail the implication of the bulge properties with different morphologies in Section 4.1.

3.2.2. Uncertainties of Best-fit Parameters

The HST observations were designed specifically to provide at least rudimentary color information with the filter combination of rest-frame B and I. In view of the possibility that the central region of the galaxy might be contaminated mildly by scattered light from the AGN (Section 3.2.1), we perform the color analysis on the images after subtracting the best-fit nucleus component, if present. We can generate color images in a straightforward manner for the six targets that were observed with the same (UVIS) detector. The majority (23/29), however, were observed with two detectors with different pixel scales and spatial resolutions. We rebin the UVIS images to match the pixel scale of the IR images, and we convolve the images taken in one filter with the corresponding PSF of the other filter. The color maps are generated by ${(B-I)}_{\mathrm{WFC}3}\,=-2.5\mathrm{log}({f}_{{B}_{\mathrm{WFC}3}}/{f}_{{I}_{\mathrm{WFC}3}})+{\mathrm{ZP}}_{{B}_{\mathrm{WFC}3}}-{\mathrm{ZP}}_{{I}_{\mathrm{WFC}3}}$, where ${f}_{{B}_{\mathrm{WFC}3}}$ and ${f}_{{I}_{\mathrm{WFC}3}}$ are the counts in B_WFC3 and I_WFC3, respectively, and ${\mathrm{ZP}}_{{B}_{\mathrm{WFC}3}}$ and ${\mathrm{ZP}}_{{I}_{\mathrm{WFC}3}}$ are the corresponding zero-points. We apply the IRAF task ellipse to the color images to derive radial color profiles.

Figure 6 gives the color maps and color profiles. For each object, the upper panel shows the B_WFC3 image, the I_WFC3 image, and the (B − I)_WFC3 color map. The color radial profile is shown in the bottom panel. We will discuss the results in Section 4.1.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We convert the I_WFC3 magnitude of the bulge (${M}_{I,\mathrm{bul}}$) to the bulge stellar mass (M_bul) using a mass-to-light ratio (M/L) inferred from the rest-frame (B − I) color. Adopting the color-based conversion from Bell & de Jong (2001),

$$\begin{eqnarray}&&\mathrm{log}\left(\displaystyle \frac{{M}_{\mathrm{bul}}}{{M}_{\odot }}\right)=-0.4({M}_{I,\mathrm{bul}}-{M}_{I,\odot })+0.439(B-I)-0.594,\end{eqnarray} \tag{ 1 }$$

where M_I,⊙ = 4.08 is the I-band absolute magnitude of the Sun (Binney & Merrifield 1998). This relation assumes a Kroupa (2001) stellar initial mass function. The stellar masses have typical errors of ∼0.1 dex due to uncertainties in stellar population (Bell et al. 2003; Conroy et al. 2009).

We use the SDSS Data Release 7 (DR7) spectrum of each object to calculate the k-correction and the color conversion from the B_WFC3 and I_WFC3 filters to conventional B and I filters. Strictly speaking, the SDSS spectrum, taken using a 3'' fiber, covers the central ∼5.6 kpc of the host galaxy (for median z ≈ 0.1), much larger than the size scale of the bulge. Nevertheless, we will use it and confirm later that the stellar mass derived from the spectrum well represents the bulge mass.

Note that the spectral range of the SDSS spectrum is not wide enough to cover the I_WFC3 bandpass. We, therefore, derive the stellar spectrum by fitting the SDSS spectrum (see Figure 7 for an example) with a Bruzual & Charlot (2003, BC03) stellar population synthesis model consisting of four stellar components with ages in the range [0.04, 0.3], [0.3, 1.0], [0.0, 3.0], and 8.0 Gyr, assuming solar metallicity. Strong emission lines are excluded from the fit. Meanwhile, Galactic extinction is removed adopting R_V = 3.1 and the Milky Way extinction law of Cardelli et al. (1989). As previously noted, contamination from nuclear scattered light is not entirely negligible for 14 of the quasars. For consistency, a power-law continuum representing the AGN component is included in the fits for these objects. The SDSS spectra of 24 targets have sufficiently high S/N (≳15) that their best-fit BC03 model yields a reduced χ² ≈ 1. For the remaining five objects with spectra of lower quality (S/N < 7), we simply adopt a model consisting of an old (10 Gyr) and a young (2.1 Gyr) population (Canalizo & Stockton 2013).

Zoom In Zoom Out Reset image size

In view of the fact that the SDSS spectrum generally covers larger regions of the host galaxy than the scale of bulge, we consider the SDSS spectrum to be representative of the whole host galaxy and derive an alternative estimate of the bulge stellar mass. We first obtain the total stellar mass of the host galaxy (M_*) using its integrated I-band magnitude and (B − I) color by Equation (1), and then we attribute a fraction B/T thereof to the bulge (M_bul,B/T). These alternative estimates of the bulge masses derived from B/T agree well with those derived from bulge magnitudes: $\langle \mathrm{log}{M}_{\mathrm{bul}}-\mathrm{log}{M}_{\mathrm{bul},B/T}\rangle \,=0.03\pm 0.16$ dex.

Another check on our stellar masses can be made using the subset of 19 objects that overlap with the MPA-JHU catalog⁸ of spectral measurements from SDSS DR7. The stellar masses in this catalog were derived from spectral energy distribution fits to the DR7 photometry⁹ The median total stellar masses from DR7 (${M}_{* ,\mathrm{DR}7}$) agree well with our estimates: $\langle \mathrm{log}{M}_{* }\,-\mathrm{log}{M}_{* ,\mathrm{DR}7}\rangle =-0.01\pm 0.13$ dex.

Table 3 lists the stellar masses derived in this work. For simplicity, in the following discussion, we simply adopt the bulge masses (M_bul) estimated from Equation (1). We note that none of the conclusions of this paper depends on which of the two bulge masses we choose.

Table 3.  Stellar Masses and Color Gradients

Object	$\mathrm{log}{M}_{* }$	$\mathrm{log}{M}_{\mathrm{bul},B/T}$	$\mathrm{log}{M}_{\mathrm{bul}}$	${\rm{\nabla }}{(B-I)}_{\mathrm{in}}$	∇(B − I)out	Morphology
	(M⊙)	(M⊙)	(M⊙)
(1)	(2)	(3)	(4)	(5)	(6)	(7)
J011935.63	10.71	10.29	10.45	0.64	−0.72	spiral (unbarred)
J074751.56	11.03	10.14	10.06	1.24	1.65	spiral (barred)
J075329.93	11.00	10.90	10.92	0.83	−2.99	merging
J075940.95	10.67	9.97	9.75	0.73	0.79	lenticular
J080252.92	11.27	11.07	11.10	0.63	−0.66	merging
J080337.32	11.30	11.20	11.16	−1.27	−3.91	lenticular
J080523.29	11.12	10.43	10.75	1.17	−0.86	merging
J081100.20	11.18	10.64	10.89	0.60	−1.65	spiral (unbarred)
J084107.06	11.03	11.03	11.03	0.64	−2.43	merging
J084344.99	11.05	10.83	10.79	0.07	−2.44	disturbed
J090754.07	10.92	10.26	10.23	0.90	1.29	spiral (barred)
J091819.66	10.26	10.26	10.26	1.70	−1.76	elliptical
J093625.36	10.28	9.73	10.11	1.15	1.17	lenticular
J103408.59	10.99	10.76	10.66	−0.10	−2.20	disturbed
J105208.19	10.65	9.81	9.97	0.20	−0.60	spiral (barred)
J110213.01	10.95	10.62	10.90	0.35	−2.57	merging
J111015.25	9.70	9.70	9.70	1.34	1.23	elliptical
J113710.77	10.91	10.91	10.91	0.39	−1.57	elliptical
J115326.42	10.56	9.82	9.79	1.02	1.48	spiral (barred)
J123804.81	10.93	9.73	10.01	1.52	0.13	spiral (barred)
J125850.77	10.96	10.12	10.05	0.75	0.82	spiral (barred)
J130038.09	10.86	10.46	10.30	0.48	0.63	lenticular
J133542.49	10.88	10.88	10.88	0.36	−1.47	disturbed
J140541.21	10.58	10.32	10.27	0.41	−0.05	spiral (unbarred)
J140712.94	10.97	10.44	10.43	1.69	1.54	spiral (barred)
J144038.09	11.01	10.14	9.94	−0.29	1.45	merging
J145019.18	11.14	11.04	11.10	0.03	−1.90	lenticular
J155829.36	10.71	9.76	9.64	0.59	1.06	spiral (barred)
J162436.40	10.99	10.46	10.38	1.21	−0.13	merging

Note. Column (1): Galaxy name. Column (2): Total stellar mass of host galaxy. Column (3): Bulge mass calculated from M_* and B/T. Column (4): Bulge mass calculated from M/L ratio of Equation (1). Column (5): Color gradients of the galaxy inner regions. The units are Δmag per dex in radius (arcsec). Column (6): Color gradients of the galaxy outer regions. Column (7): Morphological classification.

Download table as:  ASCIITypeset image

Black hole accretion and star formation are often suggested to go hand in hand (Springel et al. 2005; Hopkins et al. 2006), but the verdict from observations is somewhat mixed. While some studies of the stellar population of AGN host galaxies find nuclear and starburst activity to be broadly synchronized (e.g., Tadhunter et al. 2011; Bessiere et al. 2014, 2017), others maintain that AGN activity significantly lags behind star formation (e.g., Wild et al. 2010; Canalizo & Stockton 2013). The color information from this study provides some fresh insights into this issue, from the point of view of luminous, obscured quasars that should be experiencing both intense star formation and black hole growth.

The color maps shown in Figure 6 illustrate that in most of the hosts, the central regions of the bulge are generally bluer than their outer regions. This holds for most of the merging/disturbed hosts and most of the disk galaxies. The few that do not follow this trend may be affected by dust reddening (i.e., some of the disturbed systems and two of the lenticulars with prominent inner dust lanes). Consistent with the expected behavior of galaxies, almost all the global color profiles initially exhibit a negative color gradient in their outermost regions (become redder toward smaller radii). However, upon reaching the central regions, roughly on the scale of the effective radius of the bulge, most of the color profiles flatten or even turn over at smaller radii.

To quantify this effect, we derive the color gradients of the inner and outer regions of the host galaxies (see Table 3), and we directly compare them with a control sample of inactive galaxies measured in the same manner from the Carnegie-Irvine Galaxy Survey (CGS; Ho et al. 2011). The inner region is defined as the range from half of the PSF FWHM (to avoid possible AGN contamination, even after subtracting the nucleus) to the effective radius of the bulge (R_e); the outer region extends between R_e and 2.5R_e. Following Li et al. (2011), the color gradient, corrected for Galactic extinction, is calculated as the slope of the color profile, representing the change in color per dex in radius:

$$\begin{eqnarray}&&{\rm{\nabla }}{(\mathrm{color})}_{\mathrm{in}/\mathrm{out}}=\displaystyle \frac{{(\mathrm{color})}_{{r}_{1}}-{(\mathrm{color})}_{{r}_{0}}}{\mathrm{log}{{\boldsymbol{r}}}_{1}-\mathrm{log}{{\boldsymbol{r}}}_{0}},\end{eqnarray} \tag{ 2 }$$

where r₁ and r₀ correspond to R_e and 0.5 FWHM of PSF for the inner region, and 2.5R_e and R_e for the outer region. As in Li et al. (2011), color gradients larger than 0.1 are considered positive, between −0.1 and 0 are flat, and smaller than −0.1 are negative. A positive slope indicates that the galaxy is getting bluer toward the center.

Regardless of the sign of the outer color gradients, most of our objects exhibit positive inner B − I color gradients. Figure 8 compares the inner color gradients of our sample, split by morphological types, relative to a comparison sample drawn from CGS (Li et al. 2011). The two samples are clearly different. Apart from two S0 hosts with central dust lanes, all the early-type hosts of type 2 quasars (ellipticals and S0s) have positive inner color gradients. By comparison, 85% of the ellipticals and 64% of the S0s in CGS have flat inner color profiles. The contrast is even more pronounced for spiral galaxies, for which all spiral AGN hosts have clearly positive inner color gradients, whereas negative gradients characterize the majority (65%–88%, depending on the exact morphological type) of the bulges of inactive spirals. Even merging/disturbed systems hosting type 2 quasars, whose environments expected to be dusty, mostly exhibit positive inner color gradients. The host galaxies of type 2 quasars have a preponderance of blue central regions compared to normal galaxies of similar Hubble type, consistent with enhanced ongoing or recent star formation.

Zoom In Zoom Out Reset image size

Additionally, less direct but nevertheless compelling evidence for enhanced central star formation in type 2 quasars comes from inspection of the detailed structural parameters of the host galaxies. Figure 9 illustrates the distributions of Sérsic index (n) and bulge-to-total ratio (B/T) as a function of bulge stellar mass (M_bul) for the sample. Not unexpectedly, the merging/disturbed hosts (gray symbols) tend to have more massive (${M}_{\mathrm{bul}}\gtrsim {10}^{10.5}\,{M}_{\odot }$), more prominent (B/T ≳ 0.2) bulges with relatively large Sérsic indices (median n = 2.63), akin to classical bulges and consistent with the expectation that major mergers build classical bulges (Kormendy & Kennicutt 2004; Fisher & Drory 2008). The undisturbed early types in the sample (lenticulars and ellipticals; red and orange symbols, respectively) span a wide range in mass, but their Sérsic indices are large (median n = 4.02), again consistent with classical bulges. The late-type hosts are strikingly different. Most of the barred and unbarred spirals (blue and green symbols) have bulges of lower mass (M_bul ≲ 10^10.5 M_⊙), lower prominence (B/T ≲ 0.2), and characteristically lower Sérsic indices (n ≲ 2). They conform to the typical properties of pseudo bulges (Kormendy & Kennicutt 2004; Fisher & Drory 2008).

Zoom In Zoom Out Reset image size

The Kormendy relation (Kormendy 1977), an inverse correlation between the effective radius (R_e) of spheroids and their surface brightness (μ_e) within R_e, provides a useful empirical tool to distinguish bulge types. At a given R_e, pseudo bulges have lower μ_e than classical bulges or elliptical galaxies (Kormendy & Kennicutt 2004; Gadotti 2009; Fisher & Drory 2010). Figure 10 (top) examines the Kormendy relation of our sample using best-fit parameters from the I_WFC3 band. We fit the relation only for the classical bulges (i.e., ellipticals and the bulges of merging/disturbed and lenticular galaxies), which is denoted by the solid line. It is surprising that the late-type galaxies, which, as argued above, contain pseudo bulges, do not depart from the relation of classical bulges. This grossly deviates from the behavior of inactive galaxies, for which pseudo bulges scatter systematically below the locus of classical bulges and ellipticals. Why do pseudo bulges not appear in the Kormendy relation of type 2 quasar host galaxies? The simplest and most plausible explanation is that the bulge effective surface brightnesses have been enhanced because of excess light from recent or ongoing star formation. Together with the color gradients previously discussed, we conclude that the central regions of these obscured quasars have a characteristically young stellar population, presumably associated with a recent episode of star formation. A similar trend is also reported in the hosts of type 1 AGNs (Kim & Ho 2019).

Zoom In Zoom Out Reset image size

The role of mergers in governing AGN activity remains a vexingly controversial topic. From a theoretical point of view, major external triggers are thought to be necessary to supply the high mass accretion rates needed to sustain the luminosities of the most powerful quasars (Shlosman et al. 1990). The less stringent fueling requirements of weaker AGNs can be met through internal secular processes, such as angular momentum transported by bars (e.g., Ho et al. 1997; Somerville et al. 2008; Hopkins & Hernquist 2009). A substantial body of observational evidence broadly supports this thesis. From their analysis of a large sample of AGNs spanning a wide range of bolometric luminosities and redshifts, Treister et al. (2012) report a strong correlation between the fraction of host galaxies experiencing major mergers and AGN luminosity. A high incidence of merger features is frequently found for luminous AGNs, be they unobscured (e.g., Letawe et al. 2010; Liu et al. 2012; Hong et al. 2015), dust-reddened (Urrutia et al. 2008; Kocevski et al. 2015; Fan et al. 2016), or highly obscured (Villar-Martín et al. 2011; Bessiere et al. 2012; Wylezalek et al. 2016; Donley et al. 2018; Urbano-Mayorgas et al. 2019). But not everyone agrees. A significant number of studies of luminous AGNs at moderate (0.5 < z < 0.8; Villforth et al. 2014) and high (z ≲ 2; Schawinski et al. 2011, 2012; Mechtley et al. 2016; Villforth et al. 2017) redshifts dismiss the importance of major mergers in driving nuclear activity.

One of the most surprising results of this study is the sheer diversity of the morphologies of the host galaxies of obscured quasars. As summarized in Section 1, the traditional gas-rich major merger scenario for quasar evolution predicts that type 2 quasars should be hosted by morphologically highly disturbed galaxies. This basic expectation is not supported by our observations, at least not for a sizable fraction of our sample. Among the 29 objects studied, only 10 (34%) show clear morphological signatures of interactions or mergers, rising at most to 13 (45%) if we regard the three ellipticals as post-merger products. The majority (55%; 11 spirals and 5 lenticulars) possess normal disks. Even if we accept that lenticulars may be remnants of major mergers (Eliche-Moral et al. 2018)—by no means a universal view (Kormendy et al. 2009)—a substantial fraction (38%) are incontrovertibly ordinary unbarred or barred late-type spiral galaxies. Although it is stated that disks can survive from gas-rich major mergers under some circumstances (e.g., Barnes & Hernquist 1996; Springel & Hernquist 2005; Hopkins et al. 2009), the remnants cannot significantly fuel central black holes (Hopkins & Hernquist 2009), and the regrowth of the disk (Hopkins et al. 2009; Bundy et al. 2010) takes much longer that the typical quasar lifetime (e.g., Porciani et al. 2004; Hopkins et al. 2005; Shen et al. 2007). Therefore, the late-type host galaxies of type 2 quasars, which most likely possess pseudo bulges (Section 4.1), probably never experienced major mergers during their lifetimes. Secular processes presumably were responsible for their black hole growth. Interestingly, most of the spirals (8 out of 11) contain a bar (Figure 2).

The subset of our type 2 quasars hosted by spirals has a typical [O iii] λ5007 luminosity of 10⁹ L_⊙, not dissimilar from those hosted by earlier type galaxies or mergers. For an [O iii] bolometric correction of 600 (Kauffmann & Heckman 2009), this corresponds to a bolometric luminosity of ${L}_{\mathrm{bol}}=2.3\times {10}^{45}\,\mathrm{erg}\,{{\rm{s}}}^{-1}$, or a mass accretion rate of $\dot{M}={(\epsilon {c}^{2})}^{-1}{L}_{\mathrm{bol}}\approx 0.4\,{M}_{\odot }\,{\mathrm{yr}}^{-1}$, where c is the speed of light and  = 0.1 is the radiative efficiency. This level of fueling is modest, reflecting the fact that our sample of low-redshift obscured AGNs, although technically considered as "quasars", are, in fact, quite modest in power. Bar-driven gravitational torques may suffice to transport cold gas at this rate to the central regions of spiral galaxies (Haan et al. 2009).