Resolving the Interstellar Medium in the Nuclear Region of Two z = 5.78 Quasar Host Galaxies with ALMA

Quasars at high redshifts provide the best opportunity to study the formation and early evolution of the first supermassive black holes (SMBHs) and their host galaxies. Due to the great success of wide-field optical and near-infrared surveys in the last two decades, more than two hundred quasars have been discovered at z ≥ 5.7, an epoch close to the end of cosmic reionization (e.g., Fan et al. 2006; Jiang et al. 2009, 2016; Mortlock et al. 2011; Venemans et al. 2013; Bañados et al. 2016, 2018; Matsuoka et al. 2016, 2018b, 2018a; Chehade et al. 2018). This quasar sample covers a wide range of SMBH masses and quasar luminosities, including the most massive sources with SMBH mass above 10¹⁰ M_⊙ (e.g., Fan et al. 2000; Wu et al. 2015) as well as objects with SMBH masses of a few 10⁷–10⁸ M_⊙ (Jiang et al. 2009; Willott et al. 2010; Matsuoka et al. 2016, 2018b, 2018a) which are more similar to the common quasar population observed at lower redshifts.

The dust and gas in the host galaxies of these earliest quasars have been studied at submillimeter and millimeter wavelengths using thermal continuum emission, molecular (mostly CO), and fine structure lines, such as the [C II] 158 μm emission line (Bertoldi et al. 2003a, 2003b; Petric et al. 2003; Priddey et al. 2003; Robson et al. 2004; Walter et al. 2004, 2009; Wang et al. 2008, 2010, 2013; Riechers et al. 2009; Omont et al. 2013; Bañados et al. 2015; Willott et al. 2015, 2017; Venemans et al. 2016, 2017a, 2017b, 2017c; Decarli et al. 2017, 2018; Izumi et al. 2018). Dust continuum surveys at ∼0.5–1 mJy rms sensitivity have detected the most far-infrared (FIR)-luminous objects, with FIR luminosities of a few ×10¹² to 10¹³ L_⊙ and dust masses of >10⁸ M_⊙ (Bertoldi et al. 2003a; Priddey et al. 2003; Robson et al. 2004; Wang et al. 2007, 2008). These FIR luminous quasars are bright in CO line emission and [C II] 158 μm fine structure line emission, indicating molecular gas with masses of the order of 10¹⁰ M_⊙ and bursts of star formation in the nuclear region, coeval with the SMBH accretion and quasar activity (Bertoldi et al. 2003b; Maiolino et al. 2005; Carilli et al. 2007; Wang et al. 2013; Venemans et al. 2016; Decarli et al. 2017, 2018).

In recent years, the Atacama Large Millimeter/submillimeter Array (ALMA) has carried out comprehensive surveys of the [C II] 158 μm fine structure line in high-z quasars. For example, Decarli et al. (2018) detected [C II] in 85% of 27 optically selected quasars at z > 5.94. This line is an important coolant that traces the ionized and neutral interstellar medium (ISM) and star-forming activities (Herrera-Camus et al. 2015). The detection of [C II] line and dust continuum emission has revealed a wide range of star formation rates, from a few 10 M_⊙ yr⁻¹ to ≥1000 M_⊙ yr⁻¹ (Decarli et al. 2018; Izumi et al. 2018). The [C II]–FIR luminosity ratios of these quasar hosts range from 10⁻⁴ to a few ×10⁻³ (Maiolino et al. 2005; Wang et al. 2013; Willott et al. 2017; Decarli et al. 2018), following the trend of decreasing [C II]–FIR ratio with increasing FIR luminosities found with the IR luminous star-forming systems (Malhotra et al. 2001; Luhman et al. 2003; Hailey-Dunsheath et al. 2010; Stacey et al. 2010; Graciá-Carpio et al. 2011; Díaz-Santos et al. 2013, 2017; Muñoz & Oh 2016; Smith et al. 2017; Decarli et al. 2018; Gullberg et al. 2018; Rybak et al. 2019). In particular, the quasar hosts with high FIR luminosities of a few 10¹² to 10¹³ L_⊙ show low [C II]–FIR ratios similar to that found in the ultraluminous infrared galaxies (ULIRGs) and submillimeter galaxies (SMGs; Luhman et al. 2003; Díaz-Santos et al. 2013, 2017; Wang et al. 2013; Rybak et al. 2019). The [C II]–FIR ratio is suggested to be related to the local conditions of the ISM; e.g., a decrease of [C II]–FIR ratio could be due to a high gas temperature (i.e., T > ΔE/k ∼ 91 K, ΔE is the energy separation of the two levels of the [C II]158 μm transition) where the upper level of the [C II] transition is saturated (Muñoz & Oh 2016), or a high gas surface density (e.g., the typical density in compact ULIRGs and SMGs) where more gas is in the molecular phase (Narayanan & Krumholz 2017). It is interesting to see how the resolved distributions of the line and continuum surface brightnesses and emission ratios in these FIR luminous quasar hosts compare to that in the ULIRGs and SMGs (Smith et al. 2017; Gullberg et al. 2018; Rybak et al. 2019).

The [C II] 158 μm and molecular CO lines are bright tracers of the gas content that could be detected from distant galaxies (Solomon & Vanden Bout 2005; Carilli & Walter 2013). Measurements of the line widths, profiles, and velocity maps of these emission lines reveal important information on the gas kinematics (Walter et al. 2004; Ivison et al. 2011; Maiolino et al. 2012; Wang et al. 2013; Venemans et al. 2016, 2019). Observations of the [C II] and CO emission lines from the young quasar hosts at z ≥ 5.7 at subarcsecond to arcsecond resolution show a range of kinematic properties, including velocity gradients of ordered motion (Willott et al. 2013, 2017; Venemans et al. 2016; Shao et al. 2017; Feruglio et al. 2018), large velocity dispersion/turbulence (e.g., SDSS J231038.88+185519.7, Feruglio et al. 2018), gas outflows (e.g., SDSS J114816.64+515250.3, Maiolino et al. 2012; Cicone et al. 2015), and very compact sources with no evidence of rotation (e.g., ULAS J112001.48+064124.3, Venemans et al. 2017c). All these submillimeter/millimeter (submm/mm) observations argue for an early phase of SMBH-galaxy coevolution in the z ≥ 5.7 quasar sample. Deep imaging of the gas component at the kiloparsec scale is required to fully resolve the gas distribution and kinematic properties in these systems.

In this paper, we present the observations, results, and conclusion for two out of three quasars from our ALMA Cycle 1 program. Our team has carried out ALMA observations at 02–03 angular resolution to image the [C II] line emission from the host galaxies of three quasars, ULAS J131911.29+095051.4 (hereafter ULAS J1319+0950), SDSS J104433.04−012502.2 (hereafter SDSS J1044−0125), and SDSS J012958.51−003539.7 (hereafter SDSS J0129−0035) at z ≥ 5.7. These three objects are among the brightest dust continuum, and [C II] detections at z ∼ 6 (Wang et al. 2013), with FIR luminosities of 5 ∼ 10 × 10¹² L_⊙ and¹⁹ [C II] luminosities of 1.5 ∼ 4.4 × 10⁹ L_⊙ (Wang et al. 2013), suggesting massive star formation in the quasar hosts. The three objects are different in quasar UV luminosities and SMBH masses. SDSS J1044−0125 and ULAS J1319+0950 are among the most optically luminous quasars at z ∼ 6 with rest-frame 1450 Å apparent magnitudes of m₁₄₅₀ = 19.21 and 19.65, respectively (Fan et al. 2000; Mortlock et al. 2009). The SMBH mass is (2.7 ± 0.6) × 10⁹ M_⊙ for ULAS J1319+0950 (Shao et al. 2017) and (5.6 ± 0.6) × 10⁹ M_⊙ for SDSS J1044−0125 (Shen et al. 2019). SDSS J0129−0035 is much fainter in quasar emission with m₁₄₅₀ = 22.28 (Jiang et al. 2009). If Eddington accretion is assumed (Willott et al. 2010), the SMBH mass is 1.7 × 10⁸ M_⊙. These objects provide an ideal sample for a pilot study of the ISM distribution and dynamics in the earliest SMBH-starburst systems over a range of quasar luminosities.

We report our ALMA [C II] line imaging of ULAS J1319+0950 at 03 resolution in Shao et al. (2017). The [C II] velocity map of this object shows clear velocity gradients and the spectrum appears to be broad (FWHM ∼ 540 km s⁻¹) and flat at the top, suggesting that the [C II]-emitting gas is mainly distributed in an inclined rotating disk extending to a radius of ∼3.2 kpc. We performed dynamical modeling of the gas velocity field and measured the disk inclination angle and rotation velocity of the atomic/ionized gas in the nuclear region. The derived host galaxy dynamical mass is 1.3 × 10¹¹ M_⊙, suggesting a black hole-to-host galaxy mass ratio of ∼0.02, i.e., about four times higher than the value expected from the black hole–bulge mass relation of local galaxies (Kormendy & Ho 2013).

We here present ALMA [C II] imaging of the remaining two quasars, SDSS J0129−0035 and SDSS J1044−0125. We describe the observations in Section 2, present the results in Section 3, discuss the surface brightnesses of the [C II] and dust emission and gas kinematic properties in Section 4, and summarize the main conclusions in Section 5. We adopt a ΛCDM cosmology with H₀ = 71 km s⁻¹ Mpc⁻¹, Ω_M = 0.27, and Ω_Λ = 0.73 throughout this paper (Spergel et al. 2007).

We carried out observations of the [C II] 158 μm emission line from SDSS J0129−0035 and SDSS J1044−0125 using the Band 7 receiver on ALMA in Cycle 1 (Program ID: 2012.1.00240.S). The data was taken in 2015 July and August in the C34-6 configuration with 32 12 m diameter antennas and a maximum baseline of 1.6 km. This results in a full width at half maximum (FWHM) synthesized beam size of ∼02 using robust weighting, which corresponds to 1.2 kpc at z ∼ 5.8. We tuned the four 1.785 GHz spectral windows with one window centered on the redshifted [C II] line frequencies of the two targets and the other three windows observing the continuum. The correlator channel width is 15.625 MHz, corresponding to a velocity resolution of ∼17 km s⁻¹. We checked the phase every 5 ∼ 7 minutes on nearby calibrators, and the flux scale was calibrated on Ceres and J1058+015 with calibration uncertainties less than 10%. The total on-source time is about 80 minutes for SDSS J0129−0035 and 75 minutes for SDSS J1044−0125.

The data were calibrated using the pipeline of the Common Astronomy Software Application (CASA) Version 4.3.1. We then combined the data with that obtained from previous ALMA Cycle 0 observations (Program ID: 2011.0.00206.S, Wang et al. 2013). The ALMA Cycle 0 data alone has a lower angular resolution of 06–07, and 1σ rms sensitivities of 0.35 mJy beam⁻¹ for SDSS J0129−0035 and 0.65 mJy beam⁻¹ for SDSS J1044−0125 in each 67 km s⁻¹ channel (Wang et al. 2013). By including the Cycle 0 data, we could better sample the uv plane at different scale and improve the sensitivity at short uv-distance. As the Cycle 0 data were calibrated using a much earlier version of CASA, we corrected the weights before combining with the Cycle 1 data.²⁰ We checked the amplitudes of visibility data from the two observations at overlapping uv-distance, and the flux scales are consistent with each other within the calibration uncertainties. We averaged, in the visibility domain, the emission in line-free channels to form a "pseudo" continuum data set which we then subtracted from the uv-data to obtain a line-only data set. We made the continuum image with the data from three line-free spectral windows and made the line image cubes with the continuum-subtracted visibility data set. The images were cleaned using the CLEAN task in CASA using robust = 0.5 for weighting. The typical rms noise is 0.20 mJy beam⁻¹ per 17 km s⁻¹ channel for the final image cube of SDSS J0129−0035, and 0.36 mJy beam⁻¹ per 17 km s⁻¹ channel for SDSS J1044−0125. The rms noise of the continuum is 21 μJy beam⁻¹ for SDSS J0129−0035, and 40 μJy beam⁻¹ for SDSS J1044−0125.

For each of the two objects, we integrated the [C II] emission over the line-emitting channels to obtain velocity-integrated maps and averaged the line-free channels to obtain the continuum maps (Figure 1). We then integrated the emission within the 2σ contour region to get total [C II] line fluxes and continuum flux densities (Table 1). For SDSS J0129−0035, the [C II] velocity-integrated map has a peak of 0.80 ± 0.03 Jy beam⁻¹ km s⁻¹ (signal-to-noise ratio, S/N = 23.5), which is about 38% of the total line flux. The continuum intensity map has a peak of 1.58 ± 0.02 mJy beam⁻¹ (S/N = 80), 56% of the integrated continuum flux density. For SDSS J1044−025, the [C II] peak is 0.64 ± 0.08 Jy beam⁻¹ km s⁻¹ (i.e., S/N = 8.4). This is about 45% of the total line flux. The peak of the continuum emission is 1.88 ± 0.04 mJy beam⁻¹ (S/N = 47.5), which is about 63% of the total continuum flux density. Thus, we consider the [C II] and continuum emission from both objects are resolved.

Zoom In Zoom Out Reset image size

Table 1.  Line and Continuum Parameters

	SDSS J0129−0035	SDSS J1044−0125
[C II] line flux (Jy km s−1)	2.11 ± 0.09	1.52 ± 0.15
Peak (Jy beam−1 km s−1)	0.80 ± 0.03	0.64 ± 0.08
FWHM[C ii] (km s−1)	200 ± 8	440 ± 40
[C II] central position (J2000)	01h29m58513 −00°35'3983	10h44m33042 −01°25'0208
Deconvolved FWHM [C II] size (imfit)	(032 ± 003) × (027 ± 003)	(041 ± 007) × (030 ± 006)
FWHM [C II] size (uvmodelfit)	(032 ± 001) × (026 ± 001)	(032 ± 002) × (026 ± 003)
L[C ii](L⊙)	(1.96 ± 0.08)×109	(1.41 ± 0.14)×109
Continuum (mJy)	2.82 ± 0.04	3.02 ± 0.11
Continuum central position (J2000)	01h29m58513 −00d35'3984	10h44m33040 −01d25'0210
Deconvolved FWHM continuum size (imfit)	(018 ± 001) × (016 ± 001)	(016 ± 002) × (015 ± 002)
FWHM continuum size (uvmodelfit)	(0172 ± 0003) × (0158 ± 0005)	(0159 ± 0005) × (0154 ± 0008)
Optical quasar position (J2000)	01h29m58524 −00d35'3977	10h44m33041 −01d25'0209
Cycle 0 [C II] line fluxa (Jy km s−1)	1.99 ± 0.12	1.70 ± 0.30
Cycle 0 Continuum (mJy)	2.57 ± 0.06	3.12 ± 0.09

Note.

^aMeasurements of [C II] line flux and continuum flux density using the ALMA Cycle 0 data at 06 resolution from Wang et al. (2013). The calibration uncertainties are not included here, which is better than 15%–20% for the Cycle 0 data and better than 10% for the new data presented here.

Download table as:  ASCIITypeset image

The visibilities versus uv-distance for the [C II] line and continuum are plotted in Figure 2. The data are radially averaged with a bin size of 50 kλ for SDSS J0129−0035 and 75 kλ for SDSS J1044−0125. The decrease of the visibility amplitudes toward longer uv-distance indicates that the line and continuum sources are resolved in the uv-plane. We performed elliptical Gaussian fitting to measure the sizes of the [C II] and continuum emission in both the image plane and the uv plane, using imfit and uvmodelfit tasks in CASA, respectively. The results, together with the integrated line fluxes and continuum flux densities, are summarized in Table 1. The uv-models are also plotted in Figure 2. The Gaussian FWHM source sizes (deconvolved from the synthesized beam) and peak positions in the image plane are consistent with that from the uv-models within the fitting uncertainties. We adopt the fitting positions and deconvolved source sizes from imfit for the discussions in the rest of this paper.

Zoom In Zoom Out Reset image size

The fitting uncertainties of the [C II] peak positions are ∼7 mas in both R.A. and decl. for SDSS J0129−0035, and 25 mas for SDSS J1044−0125. The positional uncertainties for the continuum images are <3 mas for both objects. Thus, the peak positions of the [C II] line and continuum emission are consistent with each other within the fitting errors. We checked the coordinates of the optical quasars using the astrometric data from the Gaia mission (L. Lindegren et al. 2018, in preparation). For SDSS J0129−0035, the quasar coordinates are R.A. = 01^h29^m58524, and decl. = −00^d35'3977 with an uncertainty of 149 mas in R.A. and 114 mas in decl. Considering the coordinates' uncertainties, the positions of [C II] and dust continuum peaks are marginally consistent with that of the optical quasar (Figure 1). The quasar coordinates for SDSS 1044−0125 are R.A. = 10^h44^m33041, and decl. = −01^d25'0209, with an uncertainty of 27 mas in R.A. and 20 mas in decl. This is consistent with the positions of the [C II] and dust continuum peaks of this object (Figure 1).

We finally check the results by comparing the measurements to those from the low-resolution ALMA Cycle 0 data (Table 1, Wang et al. 2013). The [C II] line fluxes and continuum flux densities integrated over the source regions of both objects are consistent within the calibration uncertainties with the previous values. The ALMA data at 02 resolution reveal more detail on the spatial distribution and velocity field of the [C ii]-emitting gas component in the quasar host galaxies.

SDSS J0129−0035. We present the intensity-weighted velocity map and velocity dispersion map of the [C II] line emission in Figure 3. We adopt the [C II] redshift of z_{[C ii]} = 5.7787 as the zero velocity (Wang et al. 2013). The line is single-peaked and the velocity map shows a velocity gradient from the southeast to northwest direction. The velocity images also show two features at the edge of the [C II]-emitting region (marked as No. 1 and 2 in the velocity map, Figure 3). One has an intensity-weighted velocity of −190 km s⁻¹ at the southeast edge of the line-emitting region and the other has an intensity-weighted velocity of about +50 km s⁻¹ on the northwest part. There is no strong (≥2σ) dust emission detected in the regions of these two features. The intensity-weighted velocity dispersion map of the [C II] emission (upper right panel of Figure 3) also suggests a high velocity dispersion structure along the southeast-to-northwest direction, which seems to extend from the center to the two features. A structure with such a velocity dispersion cannot be explained with beam smearing of a rotating disk.

Zoom In Zoom Out Reset image size

The PV diagrams along the direction of the two features and the perpendicular direction are plotted in the lower panel of Figure 3. The major gas component in the PV diagrams is consistent with rotation. The two features mentioned above are presented at the regions around (−04, −190 km s⁻¹, No. 1 in Figure 3) and (+04, +50 km s⁻¹, No. 2) in the bottom left panel. It is unclear if these two components are connected to the outer part of the rotating gas disk or not. The PV diagrams also show tentative peaks (3 ∼ 4σ) with velocities of ∼ ± 200 km s⁻¹ within the central 02 (Nos. 3 and 4 in Figure 3). These components locate close to the center/peak of the [C II]-emitting region. However, the large velocities of ∼±200 km s⁻¹ do not seem to follow the velocity field of the rotating component. These tentative peaks should be checked with deeper observations at better resolution (e.g., ≲01), to confirm if they are real features and related to gas inflows/outflows.

We plot the [C II] line spectrum integrated over the line-emitting region in Figure 4, compared to the CO (6–5) line spectrum published in Wang et al. (2011). Though the CO (6–5) line profile is poorly constrained due to the low S/N, the line widths and redshifts measured with the [C II] and CO lines are consistent with each other within the uncertainties (Wang et al. 2011, 2013). Moreover, comparing to the Gaussian line profile that fit to the line spectrum, there is a tentative excess on the blue wing around −200 km s⁻¹ (arrow in Figure 4). This is likely to be related to the components seen in the velocity map and PV diagrams (e.g., Nos. 1 and 3).

Zoom In Zoom Out Reset image size

SDSS J1044−0125. We calculate the intensity-weighted velocity and velocity dispersion maps with the >4σ pixels in each [C II] channel (Figure 5), which reveals the gas kinematics in the central high surface brightness part of the line-emitting region. The gas does show some velocity structure over this area. For example, the gas component to the northwest of the quasar position shows more positive velocities. However, the velocity map seems more complicated than what one would expect from a simple rotating disk. The PV diagrams also suggest a very turbulent gas velocity field; the gas is distributed over a compact region of about 2.4 kpc (i.e., ±02) with velocities from −200 to ≳200 km s⁻¹.

Zoom In Zoom Out Reset image size

The integrated [C II] line spectrum shows multiple peaks and is not well fit by a single-Gaussian profile (Figure 6). We try to fit the line profile with three Gaussian components (A, B, and C). The fitting results of the line centers, FWHM widths, and fluxes are listed in Table 2. Here, we adopt the [C II] redshift of z_{[C ii]} = 5.7847 from the single-Gaussian fitting as the zero velocity (Wang et al. 2013). Wang et al. (2013) found a velocity offset between the Gaussian-fit line centers of the [C II] line and the CO (6–5) line for SDSS J1044−0125. With the new ALMA [C II] data in this work, we still see this offset. i.e., The CO (6–5) line overlaps only the blue part (−300–0 km s⁻¹) of the [C II] line emission with a line center at around −100 ± 44 km s⁻¹ (Wang et al. 2013) which is between the central velocities of components A and B. The mismatch between the [C II] and CO (6–5) spectra argues for complex gas components and kinematic properties in the nuclear region of SDSS J1044−0125 (see more discussion in Section 4.2).

Zoom In Zoom Out Reset image size

On the velocity-integrated map of the [C II] line (lower left of Figure 1 and left of Figure 7), there is a tentative peak 082 (i.e., 4.9 kpc) to the east of the major component. The spectrum of the emission from this secondary peak position shows a positive signal from about −740 to 400 km s⁻¹. We integrate the emission in this velocity range and show the intensity map in the left panel of Figure 7. The secondary peak is detected at about 5σ on the image, with a flux of 0.52 ± 0.10 Jy km s⁻¹. We also check the Cycle 0 data and the image at 06 resolution does show an extension toward the secondary peak location (see red contours in the upper left panel of Figure 7). A further check shows that this component is also present in a continuum-unsubtracted image in the same velocity range, as well as the image using only Cycle 1 data. No emission is detected at this location with the data from the other three continuum spectral windows. Fitting this component with a Gaussian profile yields an FWHM line width of 750 ± 190 km s⁻¹, centered at −190 ± 80 km s⁻¹. The [C II] luminosity derived with the peak intensity of this component is (4.8 ± 0.9) × 10⁸ L_⊙.

Zoom In Zoom Out Reset image size

4.1. Surface Brightnesses and Ratios of the Resolved [C II] and Dust Emission

The ALMA [C II] images of SDSS J0129−0035 and SDSS J1044−0125 probe the ISM in the nuclear regions on kiloparsec scales in these quasar-starburst systems. As we described in the previous section, the line and continuum emission are resolved from the two objects at 02 (i.e., ∼1.2 kpc) resolution. The deconvolved FWHM source sizes of the [C II] line emission are 1.8 ± 0.2 and 2.6 ± 0.5 times larger than that of the continuum for SDSS J0129−0035 and SDSS J1044−0125, respectively (based on the results from imfit, Table 1). This is consistent with the results found in previous ALMA imaging of other starburst quasar host galaxies and SMGs (Wang et al. 2013; Díaz-Santos et al. 2016; Venemans et al. 2016, 2017c; Cooke et al. 2018; Gullberg et al. 2018; Rybak et al. 2019). In most of these systems, the [C II] emission is found to be 1.3 to >2 times more extended than that of the dust continuum emission (Cooke et al. 2018; Gullberg et al. 2018).

Based on the [C II] and dust continuum intensity maps, we compare the surface brightnesses as a function of radius in the nuclear region of the two systems (Figure 8). We also include ULAS J1319+0950 in the comparison, which is another FIR luminous quasar-starburst system and spatially resolved in [C II] and dust continuum emission at a similar resolution in our ALMA Cycle 1 program (Shao et al. 2017, see Section 1). Here we scale the continuum intensity map (Figure 1) with a factor of L_FIR/S_con, where L_FIR is the FIR luminosity derived in Wang et al. (2013) assuming dust temperature of 47 K and emissivity index of β = 1.6 (Beelen et al. 2006), and S_con is the total continuum flux density close to the [C II] frequency in Table 1. This converts the monochromatic continuum map to a map of surface brightness of the FIR emission. We calculate the mean [C II] and FIR surface brightnesses within the rings at central radii of 0'', 01, 02, 03, 04, 05, and 06 and a ring width of 01. The scale of 01 corresponds to a physical size of 0.58 kpc for ULAS J1319+0950, and 0.60 kpc for SDSS J0129−0035 and SDSS J1044−0125. The symbols and error bars shown in Figure 8 represent the mean and standard deviation of the pixel values in each ring. The [C II]-to-FIR surface brightness ratios are also plotted in the right panel of Figure 8. We model the surface brightness with an exponential disk in the two-dimensional image plane with a Sérsic index of n = 1. We then convolve the model with the synthesized beam, calculate the mean surface brightnesses of the model in each ring described above, and fit these mean values to the data. The models are shown as dashed lines in Figure 8. They suggest that the [C II] and dust emission in the nuclear region of the quasar host galaxies follow an exponential light profile. The resolved surface brightness ratios between the [C II] and FIR dust emission range from 0.0001 to 0.001 within the central region where both [C II] and dust continuum are detected.

Zoom In Zoom Out Reset image size

In Figure 9, we compare the resolved [C II] and FIR continuum emission at different radii in these three z ∼ 6 quasars with other systems on the [C II]–FIR ratio versus FIR surface brightness plot. We collect samples of star-forming galaxies and quasars at low and high redshifts that have continuum source size measurements from ALMA, including the GOALS sample of local luminous infrared galaxies (Díaz-Santos et al. 2013, 2017; Shangguan et al. 2019), high-redshift SMGs (Riechers et al. 2013; Neri et al. 2014; Gullberg et al. 2018), Lyman Break Galaxies (Capak et al. 2015; Jones et al. 2017; Hashimoto et al. 2019), and the z ∼ 6–7 quasars (Wang et al. 2013; Díaz-Santos et al. 2016; Venemans et al. 2016, 2017c; Decarli et al. 2018). For the GOALS sample, we use the [C II]–FIR ratio extract from the central few kiloparsec region of the galaxies (i.e., the 94 × 94 region of one PACS detector element where the [C II] and continuum emission peaks, see details in Díaz-Santos et al. 2013, 2017). This is comparable to the [C II] and dust emission region detected in the high-z samples. We recalculate the FIR luminosity surface brightness based on the GOALs sample catalog²¹ and the decomposed infrared SED from Shangguan et al. (2019). For the high-redshift samples that have source size measurements from ALMA at 06 ∼ 07 resolution, we adopt the FIR surface brightness averaged within the half-light radius of the continuum emitting region. We also include the resolved [C II] and FIR emission from two SMGs at z = 2.9, which were imaged with ALMA at 016 resolution and have FIR luminosities and source sizes comparable to the two quasar host galaxies we studied in this work (Rybak et al. 2019).

Zoom In Zoom Out Reset image size

The continuum emission from SDSS J0129−0035 and SDSS J1044−0125, as well as ULAS J1319+0950, was resolved by ALMA on 1 ∼ 2 kpc scale and the peak FIR surface brightness is up to Σ_FIR ∼ 2 × 10¹² L_⊙ kpc⁻². This is comparable to the highest values found in the compact SMGs and ULIRGs. As was shown in Figure 9, at this FIR surface brightness, the [C II]–FIR ratio is typically an order of magnitude lower than that of normal star-forming galaxies. The [C II]–FIR ratios in the three quasar hosts get higher toward larger radii, and the FIR surface brightnesses decrease to a few 10¹⁰ L_⊙ kpc⁻². Thus the radial change of the [C II]–FIR ratios found with the three objects can be understood as a trend of decreasing [C II]–FIR ratio with increasing FIR surface brightness which was found in samples of IR luminous star-forming systems (e.g., Σ_FIR ≥ a few 10¹⁰ L_⊙ kpc⁻², Herrera-Camus et al. 2018). This trend has been termed "[C II]–FIR deficit" and was discussed in a number of papers (Hailey-Dunsheath et al. 2010; Ivison et al. 2010; Stacey et al. 2010; Farrah et al. 2013; Gullberg et al. 2015b, 2018; Díaz-Santos et al. 2016; Lutz et al. 2016; Smith et al. 2017; Cooke et al. 2018; Herrera-Camus et al. 2018).

In the compact dusty starburst systems, the high FIR surface brightness reveals high star formation rate surface density with high gas density and strong UV radiation field. As was mentioned in Section 1, in such an environment, the [C II] emission could be suppressed due to several possible mechanisms, including, e.g., saturation of the [C II] transition when the gas is heated with temperatures above 91 K, (Muñoz & Oh 2016; Rybak et al. 2019), increase of dust absorption of UV photons with high ionization parameter and a decrease of photoelectric heating efficiency (Herrera-Camus et al. 2018), or reduced amount of C⁺ medium with more carbon in the form of CO at high cloud surface density (Narayanan & Krumholz 2017). Though a model study of the origin of the [C II]–FIR deficit is beyond the goals of this work, we can conclude that the low [C II]–FIR ratio and high FIR surface brightness found in the central region of the FIR luminous quasar hosts suggest a similar high density and dusty ISM heated by a high-intensity radiation field as was found in the compact and intense starburst systems.

Of the three objects, SDSS J1044−0125 and ULAS J1319+0950 are about 10 times brighter than SDSS J0129−0035 in quasar UV luminosity (Jiang et al. 2009; Shen et al. 2019). However, their [C II]/FIR surface brightnesses and luminosity ratios are comparable. Though the central energetic active galactic nucleus (AGN) with a hard UV radiation field and strong X-ray emission may contribute to the ISM ionization and heating in the nuclear region (Stacey et al. 2010; Smith et al. 2017; Herrera-Camus et al. 2018; Shangguan et al. 2018), it is unlikely to be the dominant cause of the [C II]–FIR deficit in the quasar host galaxies (see also Decarli et al. 2018). As was pointed out by the spatially resolved studies of SMGs (Gullberg et al. 2018; Rybak et al. 2019) and local AGNs (Smith et al. 2017), the change of [C II]/FIR ratio with FIR surface brightness at different radii in the quasar host galaxies suggests that the [C II]/FIR deficit depends on the local conditions (radiation field, gas density, etc.), rather than the global properties of the systems.

We did not detect dust continuum emission at the [C II] peak to the east of the quasar SDSS J1044−0125. The continuum image sets a 3σ upper limit of 0.12 mJy for the flux density at the [C II] frequency. Assuming dust temperatures of 25 to 45 K, this constrains the FIR luminosity of this component to be <2 × 10¹¹ L_⊙, and a [C II]-to-FIR luminosity ratio of >2.4 × 10⁻³ which is much higher than the central region of the quasar host. The high [C II]-to-FIR ratio of this component reveals ISM conditions different from that in the intense starburst nuclear region of the quasar host galaxy. Deep imaging of molecular CO and other fine structure line emission will help to measure the physical properties and excitation of the medium at this offset [C II] peak and address the nature of this component (i.e., core of merging galaxy or node of outflow, etc., see more discussions below).

4.2. Gas Kinematics in the Nuclear Region

4.3. Constraints on the Host Galaxy Dynamics of SDSS J0129−0035

We here model the gas velocity map of SDSS J0129−0035 with a simple circular rotating gas disk, following the method used in Shao et al. (2017; see also Jones et al. 2017), to constrain the gas dynamics in this system. We use the ROTCUR task in the Groningen Image Processing System (GIPSY3; van der Hulst et al. 1992) to fit the observed velocity field. As was described in Shao et al. (2017), the parameters for the model are (i) the sky coordinates of the rotational center of the disk, (ii) the system velocity V_sys, (iii) the circular velocity at the fitting radius V_c(R), (iv) the position angle of the major axis on the receding half of the disk, counterclockwise from the north direction, ϕ, (v) the inclination angle between the normal direction of the ring and line of sight, incl, and (vi) the azimuthal angle related to incl, ϕ, and the disk center. We fix the disk center as the Gaussian peak center of the [C II] emission, and choose V_sys = 0 km s⁻¹ to refer to the [C II] redshift of 5.7787 (Wang et al. 2013). We then use the data within a radius of 035 to determine ϕ and incl. This is approximately the 2σ contour region of the [C II] line emission (Figure 1). This fit gives ϕ and incl as ϕ = 56 ± 1° and incl = 16 ± 20°, respectively, where the error bars denote the fitting errors. This suggests a gas disk that is very close to face-on with incl ≤36°.

In Wang et al. (2013), we adopted the same inclined circular disk assumption and estimated the inclination angle from the deconvolved axis ratio of the Cycle 0 [C II] map, i.e., incl = arccos(a_minor/a_major) = 56°. This is much larger than the value we obtained in this paper. It is likely that the two extended features (Nos. 1 and 2 in Figure 3) are unresolved in the low-resolution ALMA Cycle 0 image, which results in a larger major axis value, i.e., a_major = 041 ± 006 from the Cycle 0 data compared to a_major = 032 ± 003 in this work (Table 1). We cannot determine if these two features are part of the rotating disk or not; thus, we did not include those pixels in the fitting in this work.

Adopting the ϕ and incl values above, we constrain the circular velocities at radii of 01, 02, and 03 (Figure 10), corresponding to physical scales of 0.6, 1.2, and 1.8 kpc. If the inclination angle is fixed to 16°, the circular velocity at the largest radius (i.e., 1.8 kpc) is V_c = 250 km s⁻¹. Considering the fitting uncertainties of incl, we estimate a lower limit to the circular velocity of V_c = 110 km s⁻¹ with incl = 36°. An upper limit of the circular velocity cannot be constrained as incl could be very close to zero. Using V_c = 250 km s⁻¹, we derive the dynamical mass within a radius of 1.8 kpc to be M_dyn = 2.6 × 10¹⁰ M_⊙. If the lower limit of the circular velocity of 110 km s⁻¹ is adopted, the dynamical mass will decrease to 5.2 × 10⁹ M_⊙.

Zoom In Zoom Out Reset image size

Based on the quasar 1450 Å magnitude from the SDSS photometry (Jiang et al. 2009) and a 1450 Å bolometric correction of L_bol = 4.2 × λL_λ1450 (Runnoe et al. 2012), we derive a bolometric luminosity of 5.7 × 10¹² L_⊙ for SDSS J0129−0035. Assuming that the SMBH is accreting at the Eddington limit, we estimate the SMBH mass of this object to be ${M}_{\mathrm{BH}}={1.7}_{-0.9}^{+1.7}\times {10}^{8}\,{M}_{\odot }$. We consider an uncertainty of 0.3 dex in the SMBH mass, which represents the scatter of the Eddington ratio of the z ∼ 6 quasar sample (De Rosa et al. 2011). This results in an SMBH to host galaxy dynamical mass ratio of M_BH/M_dyn = 0.007 with V_c = 250 km s⁻¹. If the lower limit of V_c = 110 km s⁻¹ is adopted, the mass ratio will become 0.03. Considering the facts that (i) the model has large uncertainties for face-on systems and cannot accurately determine the inclination angle and circular velocity, (ii) we do not reach the flat part of the rotation curve as shown in Figure 10, and (iii) the outer part of the disk may not be traced by the [C II] velocity map due to low S/N or because it is disrupted (Lupi et al. 2019; Venemans et al. 2019), the M_dyn value derived above should be considered as a lower limit of the host galaxy dynamical mass for SDSS J0129−0035. The corresponding M_BH/M_dyn ratio hence represents an upper limit.

We compare the SMBH mass and the lower limit of the host galaxy dynamical mass of SDSS J0129−0035 to those of other quasars at z ∼ 6–7 quasars and the relationship of local galaxies (Kormendy & Ho 2013). We collect the dynamical mass measurements of the z ∼ 6–7 quasars from the literature. These objects are all detected in [C II] and/or CO line emission (Wang et al. 2013, 2016, 2019; Cicone et al. 2015; Willott et al. 2015, 2017; Venemans et al. 2016; Decarli et al. 2018; Feruglio et al. 2018; Izumi et al. 2018) and have source size measurements from [C II] or CO data and inclination angles estimated from the minor and major axis ratios of the line emission. The two objects, SDSS J0129−0035 (this work) and ULAS J1319+0950 (Shao et al. 2017), were observed with ALMA at a much better resolution of 02 ∼ 03. The [C II] emission is resolved and the velocity maps show clear velocity gradients. These, combined with the disk dynamical model described above, provide more reliable constraints on the disk inclination angle, circular velocity, and dynamical mass of the quasar host galaxies. We denote these two objects as blue and red stars in the plot. According to Figure 11, the black hole-to-host mass relationship of local galaxies given by Kormendy & Ho (2013) predicts a mass ratio of 0.0043 for an SMBH mass of 1.7 × 10⁸ M_⊙. Thus, it is unlikely that the black hole-to-host mass ratio of SDSS J0129−0035 is significantly above the local relationship. This is unlike the most luminous/massive quasars at this redshift, which usually show ratios a few to ≳10 times higher (Venemans et al. 2016; Decarli et al. 2018; Wang et al. 2019). This agrees with the results of other [C II]-detected z ≳ 6 quasars with similar or lower SMBH masses, in which the ratios are consistent with the trend defined by the local systems (Willott et al. 2010, 2017; Izumi et al. 2018). The dynamical mass constraints based on the [C II] or CO imaging of these z ∼ 6–7 quasars suggest that, in the early universe, the most massive SMBHs with masses of 10⁹–10¹⁰ M_⊙ may grow faster than that of their host galaxies (Walter et al. 2004; Wang et al. 2016, 2019; Venemans et al. 2016; Decarli et al. 2018) while the less massive systems (10⁷ ∼ 10⁸ M_⊙) are evolving more closely to the trend of local galaxies (Willott et al. 2017; Izumi et al. 2018).

Zoom In Zoom Out Reset image size

Based on new ALMA observations we present images of the [C II] and dust continuum emission from the host galaxies of two z = 5.78 quasars, SDSS J0129−0035 and SDSS J1044−0125 on 02 scales, resolving the ISM emission in the central ∼4 kpc of the hosts. The main conclusions are summarized as follows:

• 1.  
  The FWHM sizes of the [C II] emission is 1.8–2.4 kpc and that of the dust continuum ∼1kpc for the two objects, constraining the size of the intense star-forming region around the accreting SMBHs. As was found in other starburst quasar host galaxies and SMGs at high redshift, the [C II] emission regions appear to be more extended compared to that of dust continuum (Cooke et al. 2018; Gullberg et al. 2018).
• 2.  
  We derive the [C II] and FIR surface brightness and emission ratios at different radii based on the resolved [C II] and continuum images. The resolved [C II]–FIR ratios are decreasing with increasing Σ_FIR toward the center, following the trend of the IR luminous star-forming galaxies with Σ_FIR in the same range of a few 10¹⁰ L_⊙ kpc⁻² to 2 × 10¹² L_⊙ kpc⁻² (Stacey et al. 2010; Díaz-Santos et al. 2013; Gullberg et al. 2015b, 2018; Lutz et al. 2016; Cooke et al. 2018; Herrera-Camus et al. 2018; Rybak et al. 2019). The low [C II]–FIR ratios of 0.0001 ∼ 0.0002 found at the peak Σ_FIR in the central region of these two quasar hosts suggest dense ISM with high-intensity radiation field similar to that in the compact dusty starburst systems (Díaz-Santos et al. 2017; Gullberg et al. 2018; Herrera-Camus et al. 2018; Rybak et al. 2019).
• 3.  
  The velocity field and PV diagrams of SDSS J0129−0035 reveal a combination of rotation with turbulent gas clumps; We detect clear velocity gradients in the velocity map. We also see features with velocity offsets of −190 and +50 km s⁻¹ at the outer part of the [C II] emitting region and gas clumps with velocity offsets of ±200 km s⁻¹ within the central 02 region, suggesting other kinematic activities, such as inflow or outflows.
• 4.  
  The [C II]-emitting gas in the central region of SDSS J1044−0125 does not show a clear sign of rotation. The PV-diagram of the [C II] emission suggests very turbulent gas kinematics in the central area. The [C II] line spectrum shows multiple peaks, which is broader and slightly offset from the CO (6–5) line emission. In addition, we tentatively detect an offset peak, 4.9 kpc away to the east of the quasar with a very broad line width (750 ± 190 km s⁻¹) which may relate to an infalling companion source or node of outflows.
• 5.  
  Overall, the gas kinematics on kiloparsec scales in the nuclear region of the two starburst quasar host galaxies reveal significant dynamical evolution of the ISM, which could be powered by both AGN and the intense star formation.
• 6.  
  Based on the dynamical modeling of the [C II] velocity gradients of SDSS J0129−0035, we derive a disk inclination angle very close to face-on (i.e., incl = 16° ± 20°). We can thus only put a lower limit on the host dynamical mass within the [C II]-emitting region. The result argues against an SMBH-bulge mass ratio that is significantly above (e.g., larger by a factor of 10) the value derived from the local relationship. As was discussed in previous studies of the less luminous/massive quasar population at z ∼ 6, the mass ratio of SDSS J0129−0035 is likely to be closer to the local relationship than to that found for the most luminous and massive quasars at the same redshift (Willott et al. 2010; Venemans et al. 2016; Izumi et al. 2018; Wang et al. 2019).

The ALMA images of the [C II] emission in these young quasar host galaxies at kiloparsec resolution reveal rich features of gas kinematics in the nuclear region around the powerful AGNs. Between the two objects studied in this work, we do see more turbulent/clumpy gas in the nuclear region and a curious discrepancy between the [C II] and CO line profiles in SDSS J1044−0125, which hosts a more massive SMBH with higher quasar luminosity. However, it is still unclear if this is due to the feedback of the powerful AGN or nuclear star formation. If the east peak is associated with an infalling companion, the tidal perturbations could also affect the gas dynamic in the nuclear region. Deeper images of the [C II] and molecular CO emission from this object are still required to confirm these features. The high-resolution imaging studies should also be expanded to a larger sample to check for significant evolution of gas kinematic properties with AGN activity, e.g., more turbulent or disrupting gas disk in more luminous AGNs. Deep imaging of the molecular gas (e.g., with ALMA) and the stellar component (e.g., with the JWST in the future) on similar scales are also urgently required to fully probe host galaxy evolution of these young quasars at the earliest epochs.

This work was supported by National Key Program for Science and Technology Research and Development (grant 2016YFA0400703). We are thankful for the support from the National Science Foundation of China (NSFC) grant Nos. 11721303, 11373008, 11533001, 11473004, the Strategic Priority Research Program "The Emergence of Cosmological Structures" of the Chinese Academy of Sciences, grant No. XDB09000000, and the National Key Basic Research Program of China 2014CB845700. R.W. acknowledges support from the Thousand Youth Talents Program of China and the NSFC grant Nos. 11443002 and 11473004. D.R. acknowledges support from the National Science Foundation under grant No. AST-1614213. D.N. was funded in part by NSF AST-1715206 and HST AR-15043.0001. The National Radio Astronomy Observatory (NRAO) is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. This paper makes use of the following ALMA data: ADS/JAO.ALMA#2011.0.00206.S and #2012.1.00240.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO, and NAOJ. R.W. thanks Dr Matus Rybak for providing the resolved [C II] and FIR continuum measurements of their SMG sample. R.W. thanks Dr Shangguan for providing FIR luminosities of the GOALS sample.

Facilities: ALMA.