A Wide and Deep Exploration of Radio Galaxies with Subaru HSC (WERGS). II. Physical Properties derived from the SED Fitting with Optical, Infrared, and Radio Data

TL;DR

This study analyzes the physical properties of radio galaxies using multi-wavelength data and SED fitting, revealing dependencies on redshift and identifying diverse galaxy characteristics that challenge traditional views.

Contribution

First comprehensive SED-based analysis of radio galaxies combining optical, IR, and radio data, uncovering new insights into their physical properties and diversity.

Findings

Optically-faint RGs tend to be high redshift with high SFR and Eddington ratios.

Certain RGs exhibit properties differing from classical models, indicating diverse evolutionary stages.

Physical properties like $E(B-V)_*$, SFR, and IR luminosity depend on redshift.

Abstract

We present physical properties of radio galaxies (RGs) with $f_{\rm 1.4 GHz} >$ 1 mJy discovered by Subaru Hyper Supreme-Cam (HSC) and VLA Faint Images of the Radio Sky at Twenty-Centimeters (FIRST) survey. For 1056 FIRST RGs at $0 < z \leq 1.7$ with HSC counterparts in about 100 deg$^2$, we compiled multi-wavelength data of optical, near-infrared (IR), mid-IR, far-IR, and radio (150 MHz). We derived their color excess ($E (B-V)_{*}$), stellar mass, star formation rate (SFR), IR luminosity, the ratio of IR and radio luminosity ($q_{\rm IR}$), and radio spectral index ($\alpha_{\rm radio}$) that are derived from the SED fitting with CIGALE. We also estimated Eddington ratio based on stellar mass and integration of the best-fit SEDs of AGN component. We found that $E (B-V)_{*}$, SFR, and IR luminosity clearly depend on redshift while stellar mass, $q_{\rm IR}$, and $\alpha_{\rm radio}$ do not significantly depend on redshift. Since optically-faint ($i_{\rm AB} \geq 21.3$) RGs that are newly discovered by our RG survey tend to be high redshift, they tend to not only have a large dust extinction and low stellar mass but also have high SFR and AGN luminosity, high IR luminosity, and high Eddington ratio compared to optically-bright ones. The physical properties of a fraction of RGs in our sample seem to differ from a classical view of RGs with massive stellar mass, low SFR, and low Eddington ratio, demonstrating that our RG survey with HSC and FIRST provides us curious RGs among entire RG population.

Tables

Table 1. Table 1: Parameter ranges used in the SED fitting with CIGALE .

Paramerer	Value
Double exp. SFH
${\tau }_{\mathrm{main}}$ [Myr]	1000, 3000, 4000, 6000
${\tau }_{\mathrm{burst}}$ [Myr]	3, 5, 8, 15, 80
$f_{\mathrm{burst}}$	0.001, 0.1, 0.3
age [Myr]	1000, 4000, 6000, 8000, 10000
SSP (Bruzual & Charlot, 2003)
IMF	Chabrier (2003)
Metallicity	0.02
Dust attenuation (Calzetti et al., 2000)
	0.01, 0.1, 0.15, 0.2, 0.25,
$E{\left(B-V\right)}_{\ast }$	0.3, 0.35, 0.4, 0.45, 0.5,
	0.55, 0.6, 0.8, 1.0
AGN emission (Fritz et al., 2006)
$R_{\max }/R_{\min }$	60
${\tau }_{9.7}$	6.0
$\beta$	-0.50
$\gamma$	0.0
$\theta$	100.0
$\psi$	0.001, 60.100, 89.990
$f_{\mathrm{AGN}}$	0.1, 0.5, 0.9
Dust emission (Dale et al., 2014)
IR power-law slope (${\alpha }_{\mathrm{dust}}$)	0.0625, 0.2500, 1.0000, 2.0000
Radio synchrotron emission
$L_{\mathrm{FIR}}/L_{\mathrm{radio}}$ coefficient ($q_{\mathrm{IR}}$)	00.01, 0.1, 0.3, 0.5, 1.0, 2.5
spectral index (${\alpha }_{\mathrm{radio}})$	0.5, 0.7, 0.9 1.1, 1.3

Table 2. Table 2: Summary of physical properties (the mean value of each quantity) of 835 HSC–FIRST RGs.

Physical properties	SDSS-level	HSC-level	Total
$E{\left(B-V\right)}_{\ast }$	0.19	0.45	0.30
$\log \left(M_{\ast }/M_{☉}\right)$	11.26	11.08	11.19
$\log$ SFR [$M_{\ast }$ yr-1]	0.55	1.51	0.93
$\log [L_{\mathrm{IR}}$(AGN)/$L_{☉}]$	10.56	11.32	10.87
$\log \left(L_{\mathrm{IR}}/L_{☉}\right)$	11.31	12.04	11.61
${\alpha }_{\mathrm{radio}}$	0.72	0.74	0.73
$q_{\mathrm{IR}}$	0.37	0.31	0.34
$\log {\lambda }_{\mathrm{Edd}}$	-1.95	-0.94	-1.54

Table 3. Table 3: Format and column descriptions of HSC–FIRST RG catalog.

Column name	Format	Unit	Description
ID	LONG		unique id
Name	STRING		object name in the FIRST catalog
R.A.	DOUBLE	degree	Right Assignation (J2000.0) from HSC S16a wide catalog
Decl.	DOUBLE	degree	Declination (J2000.0) from HSC S16a wide catalog
Redshift	DOUBLE		Redshift
Ref_redshift	STRING		Reference of redshift (mizuki/SDSS-DR12/GAMA-DR2/WIGGLEZ-DR2)
$u$mag	DOUBLE	AB mag.	$u$-band magnitude from KiDS DR3
$u$mag_err	DOUBLE	AB mag.	$u$-band magnitude error from KiDS DR3
$g$mag	DOUBLE	AB mag.	$g$-band magnitude from HSC S16a wide catalog
$g$mag_err	DOUBLE	AB mag.	$g$-band magnitude error from HSC S16a wide catalog
$r$mag	DOUBLE	AB mag.	$r$-band magnitude from HSC S16a wide catalog
$r$mag_err	DOUBLE	AB mag.	$r$-band magnitude error from HSC S16a wide catalog
$i$mag	DOUBLE	AB mag.	$i$-band magnitude from HSC S16a wide catalog
$i$mag_err	DOUBLE	AB mag.	$i$-band magnitude error from HSC S16a wide catalog
$z$mag	DOUBLE	AB mag.	$z$-band magnitude from HSC S16a wide catalog
$z$mag_err	DOUBLE	AB mag.	$z$-band magnitude error from HSC S16a wide catalog
$y$mag	DOUBLE	AB mag.	$y$-band magnitude from HSC S16a wide catalog
$y$mag_err	DOUBLE	AB mag.	$y$-band magnitude error from HSC S16a wide catalog
$j$mag	DOUBLE	AB mag.	$j$-band magnitude from VIKING DR3
$j$mag_err	DOUBLE	AB mag.	$j$-band magnitude error from VIKING DR3
$h$mag	DOUBLE	AB mag.	$h$-band magnitude from VIKING DR3
$h$mag_err	DOUBLE	AB mag.	$h$-band magnitude error from VIKING DR3
$k$smag	DOUBLE	AB mag.	$k$s-band magnitude from VIKING DR3
$k$smag_err	DOUBLE	AB mag.	$k$s-band magnitude error from VIKING DR3
w1mag	DOUBLE	Vega mag	3.4 $\mathrm{\mu m}$ magnitude from ALLWISE
w1mag_err	DOUBLE	Vega mag	3.4 $\mathrm{\mu m}$ magnitude error from ALLWISE
w2mag	DOUBLE	Vega mag	4.6 $\mathrm{\mu m}$ magnitude from ALLWISE
w2mag_err	DOUBLE	Vega mag	4.6 $\mathrm{\mu m}$ magnitude error from ALLWISE
w3mag	DOUBLE	Vega mag	12 $\mathrm{\mu m}$ magnitude from ALLWISE
w3mag_err	DOUBLE	Vega mag	12 $\mathrm{\mu m}$ magnitude error from ALLWISE
w4mag	DOUBLE	Vega mag	22 $\mathrm{\mu m}$ magnitude from ALLWISE
w4mag_err	DOUBLE	Vega mag	22 $\mathrm{\mu m}$ magnitude error from ALLWISE
A_u	DOUBLE	mag	Galactic extinction correction for $u$-band
A_g	DOUBLE	mag	Galactic extinction correction for $g$-band
A_r	DOUBLE	mag	Galactic extinction correction for $r$-band
A_i	DOUBLE	mag	Galactic extinction correction for $i$-band
A_z	DOUBLE	mag	Galactic extinction correction for $z$-band
A_y	DOUBLE	mag	Galactic extinction correction for $y$-band
A_j	DOUBLE	mag	Galactic extinction correction for $j$-band
A_h	DOUBLE	mag	Galactic extinction correction for $h$-band
A_ks	DOUBLE	mag	Galactic extinction correction for $k$s-band
Flux_34	DOUBLE	mJy	Flux density at 3.4 $\mathrm{\mu m}$
Flux_34_err	DOUBLE	mJy	Uncertainty of flux density at 3.4 $\mathrm{\mu m}$
Flux_46	DOUBLE	mJy	Flux density at 4.6 $\mathrm{\mu m}$
Flux_46_err	DOUBLE	mJy	Uncertainty of flux density at 4.6 $\mathrm{\mu m}$
Flux_12	DOUBLE	mJy	Flux density at 12 $\mathrm{\mu m}$
Flux_12_err	DOUBLE	mJy	Uncertainty of flux density at 12 $\mathrm{\mu m}$
Flux_22	DOUBLE	mJy	Flux density at 22 $\mathrm{\mu m}$ flux density
Flux_22_err	DOUBLE	mJy	Uncertainty of flux density at 22 $\mathrm{\mu m}$
Flux_100	DOUBLE	mJy	Flux density at 100 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_100_err	DOUBLE	mJy	Uncertainty of flux density at 100 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_160	DOUBLE	mJy	Flux density at 160 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_160_err	DOUBLE	mJy	Uncertainty of flux density at 160 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_250	DOUBLE	mJy	Flux density at 250 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_250_err	DOUBLE	mJy	Uncertainty of flux density at 250 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_350	DOUBLE	mJy	Flux density at 350 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_350_err	DOUBLE	mJy	Uncertainty of flux density at 350 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_500	DOUBLE	mJy	Flux density at 500 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_500_err	DOUBLE	mJy	Uncertainty of flux density at 500 $\mathrm{\mu m}$ from H-ATLAS DR1
Flux_14G	DOUBLE	mJy	Flux density at 1.4 GHz from FIRST
Flux_14G_err	DOUBLE	mJy	Uncertainty pf flux density at 1.4 GHz from FIRST
Flux_150M	DOUBLE	mJy	Flux density at 150 MHz from TGSS ADR1
Flux_150M_err	DOUBLE	mJy	Uncertainty pf flux density at 150 MHz from TGSS ADR1
Flag_u	INT		Flag for $u$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_g	INT		Flag for $g$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_r	INT		Flag for $r$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_i	INT		Flag for $i$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_z	INT		Flag for $z$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_y	INT		Flag for $y$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_j	INT		Flag for $j$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_h	INT		Flag for $h$-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_ks	INT		Flag for $k$s-band data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_34	INT		Flag for 3.4 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_46	INT		Flag for 4.6 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_12	INT		Flag for 12 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_22	INT		Flag for 22 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_100	INT		Flag for 100 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_160	INT		Flag for 160 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_250	INT		Flag for 250 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_350	INT		Flag for 350 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_500	INT		Flag for 500 $\mathrm{\mu m}$ data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_14G	INT		Flag for 1.4 GHz data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
Flag_150M	INT		Flag for 150 MHz data (0: CIGALE. 1: CIGALE with 3$\sigma$ upper limit, 2: non CIGALE
log_L14GaaIf an object has ${\alpha }_{\mathrm{radio}}$, $\log L_{1.4\mathrm{GHz}}$ is derived from Equation 3. Otherwise, we assume ${\alpha }_{\mathrm{radio}}$ = 0.7 for a calculation (see Section 3.3).	DOUBLE	W Hz-1	Rest-frame luminosity density at 1.4 GHz
log_L14G_err	DOUBLE	W Hz-1	Uncertainty of rest-frame luminosity density at 1.4 GHz
E_BV	DOUBLE		Color excess ($E\left(B-V\right)$) derived from CIGALE
E_BV_err	DOUBLE		Uncertainty of color excess ($E\left(B-V\right)$) derived from CIGALE
log_M	DOUBLE	$M_{☉}$	Stellar mass derived from CIGALE
log_M_err	DOUBLE	$M_{☉}$	Uncertainty of stellar mass derived from CIGALE
log_SFR	DOUBLE	$M_{☉}$ yr-1	SFR derived from CIGALE
log_SFR_err	DOUBLE	$M_{☉}$ yr-1	Uncertainty of SFR derived from CIGALE
log_SFR_IR	DOUBLE	$M_{☉}$ yr-1	SFR derived from Equation 8
log_LIR	DOUBLE	$L_{☉}$	IR luminosity derived from CIGALE
log_LIR_err	DOUBLE	$L_{☉}$	Uncertainty of IR luminosity derived from CIGALE
log_LIR_AGN	DOUBLE	$L_{☉}$	IR luminosity contributed from AGN derived from CIGALE
log_LIR_AGN_err	DOUBLE	$L_{☉}$	Uncertainty of IR luminosity contributed from AGN derived from CIGALE
alpha_radio	DOUBLE		Radio spectral index (${\alpha }_{\mathrm{radio}}$) derived from Equation 1
alpha_radio_err	DOUBLE		Uncertainty of radio spectral index (${\alpha }_{\mathrm{radio}}$)
qir	DOUBLE		$q_{\mathrm{IR}}$ derived from Equation 4
qir_err	DOUBLE		Uncertainty of $q_{\mathrm{IR}}$
DOF	INT		Degree of freedom for the SED fitting
rechi2	DOUBLE		Reduced ${\chi }^2$ derived from CIGALE
log_MBH	DOUBLE	$M_{☉}$	Black hole mass derived from stellar mass
log_Lbol	DOUBLE	$L_{☉}$	Bolometric luminosity derived from the best-fit SED
log_ledd	DOUBLE		Eddington radio (${\lambda }_{\mathrm{Edd}}$)

Table 4. Table 4: Best-fit SED template of each HSC–FIRST RG.

Column name	Format	Unit	Description
ID	LONG		unique id
Wavelength	DOUBLE	$\mathrm{\mu m}$	wavelength
FNU	DOUBLE	mJy	flux density at each wavelength
LNU	DOUBLE	W	luminosity density at each wavelength

Equations

$${\alpha }_{\mathrm{radio}}=\frac{\log (F_{150\mathrm{MHz}}/F_{1.4\mathrm{GHz}})}{\log ({\nu }_{1.4\mathrm{GHz}}/{\nu }_{150\mathrm{MHz}})}$$

$$\alpha =\frac{{\sum }_i\frac{f_i{m}_i}{{\sigma }_i^2}}{{\sum }_i\frac{m_i^2}{{\sigma }_i^2}}+\frac{{\sum }_j\frac{f_j{m}_j}{{\sigma }_j^2}}{{\sum }_j\frac{m_j^2}{{\sigma }_j^2}},$$

$$L_{1.4\mathrm{GHz}}=\frac{4\pi d_L^2{F}_{1.4\mathrm{GHz}}}{(1+z{)}^{1-{\alpha }_{\mathrm{radio}}}},$$

$$q_{\mathrm{IR}}=\log \left(\frac{L_{\mathrm{IR}}/3.75\times 10^{12}}{L_{1.4\mathrm{GHz}}}\right),$$

$${\mathrm{SFR}}_{[\mathrm{OII}]}=(6.58\pm 1.65)\times 10^{-42}{L}_{[\mathrm{OII}]}^{\mathrm{cor}},$$

$$L_{[\mathrm{OII}]}^{\mathrm{cor}}=L_{[\mathrm{OII}]}^{\mathrm{obs}}10^{0.4k_{[\mathrm{OII}]}E(B-V{)}_{\mathrm{gas}}},$$

$$q_{\mathrm{excess}}=\log \left(\frac{L_{1.4\mathrm{GHz}}}{\mathrm{SFR}(\mathrm{IR})}\right),$$

$$\log \mathrm{SFR}(\mathrm{IR})=\log L_{\mathrm{IR}}(\mathrm{SF})-\mathrm{9.966.}$$

$$\log (M_{\mathrm{BH}}/M_{\text{sun}})=7.45+1.05\times \log (M_{\ast }/10^{11}{M}_{\text{sun}}),$$

Full text

infrared: galaxies — radio continuum: galaxies — catalogs — methods: observational — methods: statistical

††facilities: Subaru (HSC), VST, ESO:VISTA, WISE, Herschel, VLA, GMRT, IRSA††software: IDL, IDL Astronomy User’s Library (Landsman, 1993), TOPCAT (Taylor, 2006), CIGALE (Boquien et al., 2019)

1 Introduction

In the last decade, observational and theoretical works have reported that feedback from radio active galactic nuclei (AGNs) harbored in radio galaxies (RGs) and radio-loud quasars can play an important role in the formation and evolution of galaxies (e.g., Croton et al., 2006; Fabian, 2012). Mechanical injection of energy from RGs provides an impact on the gas reservoirs in galaxies and galaxy clusters (Morganti et al., 2013). Such AGN feedback could regulate star formation (SF) and even the growth of supermassive black holes (SMBHs) in galaxies. Therefore, it is important to investigate the physical properties related to SF and AGN activity for RGs as a function of redshift in order to understand a full picture of the formation and evolution of galaxies.

Multi-wavelength dataset of optical and infrared (IR) for RGs is crucial for studying their physical properties such as stellar mass, AGN/SF activity, and star formation rate (SFR). For example, a combination of National Radio Astronomy Observatory (NRAO) Very Large Array (VLA) Sky Survey (NVSS; Condon, 1989) or the VLA Faint Images of the Radio Sky at Twenty-Centimeters survey (FIRST; Becker et al., 1995), and the Sloan Digital Sky Survey (SDSS; York et al., 2000) provided a lager number of RGs with optical counterparts in the local Universe (Ivezić et al., 2002; Best et al., 2005; Helfand et al., 2015), allowing us a statical investigation of those “optically bright” RGs with $r<22.2$ mag at redshift $z<0.5$. These objects have been well studied in terms of UV/optical properties (e.g., de Ruiter et al., 2015), morphologies (e.g., Liske et al., 2015; Aniyan & Thorat, 2017; Lukic et al., 2018), mid-IR (MIR) properties (e.g., Gürkan et al., 2014), and far-IR (FIR) properties (e.g., Gürkan et al., 2015, 2018) as well as black hole (BH) mass and its accretion rate (e.g., Best & Heckman, 2012). Almost all of the optically bright local RGs have elliptical hosts with stellar mass of $>10^{10.5}$ M*☉* and SFR of $<$ 10 M*☉* yr*-1* (Best & Heckman, 2012). Only a small fraction of the local RGs has relatively small stellar mass with moderate star-forming activities (Smolčić, 2009; Best & Heckman, 2012).

At the high-$z$ Universe ($z>1$), known RGs are powerful or radio-luminous ($L_{\rm radio}\gtrsim 10^{26}$ W Hz*-1*, corresponding to $>0.1$ mJy). The powerful high-$z$ RGs are dominated by the evolved stellar populations with a stellar mass of $10^{11-12}~{}M_{\sun}$ (e.g., Rocca-Volmerange et al., 2004; Seymour et al., 2007; Casey et al., 2009). The IR luminosity ($L_{\rm IR}$) of those powerful high-$z$ RGs often exceed $10^{12}\,L_{\sun}$ that is classified as ultraluminous IR galaxy (ULIRG; Sanders & Mirabel, 1996). They also show the evidence of high SFR and high BH accretion rate through IR and sub-millimeter observations (e.g., Chapman et al., 2010; Magnelli et al., 2010; Seymour et al., 2012; Drouart et al., 2014; Bonzini et al., 2015). On the other hand, Falkendal et al. (2019) investigated the SFR of those powerful high-$z$ RGs based on multi-wavelength SEDs with taking into account their synchrotron emission. They reported that their SFRs are indeed lower than those of a main sequence of galaxies, suggesting an importance of multi-wavelength analysis for RGs.

Deep radio and optical observations enable us to find much more fainter RGs (see Padovani, 2016, and references therein) and to provide a comprehensive understanding by connecting RGs between local and high-$z$ Universe. Delvecchio et al. (2018) investigated RGs in the VLA-COSMOS field (Smolčić et al., 2017b) based on a multi-wavelength dataset (Smolčić et al., 2017a; Laigle et al., 2016). They found that an average BH mass accretion rate, represented by a ratio of bolometric luminosity to stellar mass, increases with increasing redshift up to $z\sim 4$. They also reported that this trend is similar to a fact that fraction of star-forming host galaxies also increases with increasing redshift. Although their statistical experiment was performed with a relatively small area ($\sim 2$ deg2), a wide-field survey with deep radio and optical facilities enables to find a large number of “optically faint” RGs, providing us a laboratory to investigate their evolution in more high resolutions in redshifts and luminosities.

Recently, Yamashita et al. (2018, Paper I) performed a systematic search for RGs and quasars as a project, so-called “the Wide and Deep Exploration of Radio Galaxies with Subaru HSC (WERGS).” They reported the result of optical identifications of radio sources detected by FIRST with the Hyper Suprime-Cam (HSC; Miyazaki et al., 2012, 2018) (see also Furusawa et al., 2018; Kawanomoto et al., 2018; Komiyama et al., 2018) Subaru Strategic Program survey (HSC-SSP; Aihara et al., 2018a). By cross-matching the final data release of the FIRST survey (Helfand et al., 2015) with HSC S16A data (Aihara et al., 2018b), they found 3579 optical counterparts of FIRST sources in a 154 deg2 of a HSC-SSP Wide field (see Section 2.1). Their radio flux densities at 1.4 GHz (20 cm) are above 1 mJy while about 60% of them are optically-faint ones with $i\geq$ 21.3 mag that are undetected by the SDSS, allowing us to explore a new parameter space, i.e., optically-faint bright radio sources. Plenty of RG and quasar sample also gives an opportunity to discover a specially rare population, for example, a RG at high redshift (Yamashita et al. in preparation) and extremely radio-loud quasars (Ichikawa et al. in preparation).

This is the second in a series of papers from the WERGS project, in which we report the physical properties of radio-loud galaxies at $0<z\leq 1.7$ with $i$-band magnitude between 18 and 26, that are derived from the Spectral Energy Distribution (SED) fitting of multi-wavelength data. In this paper, we follow the same definition of RGs and quasars as adopted in Yamashita et al. (2018). But we removed stellar objects, i.e., radio-loud quasars that are optically unresolved objects based on optical morphology (see Yamashita et al., 2018), and focus only on RGs that have optically resolved morphologies.

The structure of this paper is as follows. Section 2 describes the sample selection of RGs, the multi-wavelength dataset, and our SED modeling. In Section 3, we report the result of SED fitting and the derived physical quantities of RGs detected by the HSC and FIRST. In Section 4, we discuss a possible selection bias, an uncertainty of our SED fitting, and BH mass accretion rate for our sample. We summarize this work in Section 5. All information about our RG sample such as coordinates, multi-band photometry, derived physical quantities are available as a catalog (see Appendix A). We also provide best-fit SED templates of those RGs (see Appendix B). Throughout this paper, the adopted cosmology is a flat universe with $H_{0}$ = 70 km s*-1* Mpc*-1*, $\Omega_{\rm M}$ = 0.27, and $\Omega_{\Lambda}$ = 0.73, that are same as those adopted inYamashita et al. (2018). Unless otherwise noted, all magnitudes refer to the AB system.

2 Data and analysis

2.1 Sample selection

Figure 1 shows a flow chart of our sample selection process. The original sample was drawn from 3579 RGs and quasars in (Yamashita et al., 2018) who used the HSC-SSP and FIRST data. The HSC–SSP is an on-going optical imaging survey with five broadband filters ($g$-, $r$-, $i$-, $z$-, and $y$-band) and four narrowband filters (see Aihara et al., 2018a; Bosch et al., 2018; Coupon et al., 2018; Huang et al., 2018). This survey consists of three layers: Wide, Deep, and UltraDeep, and this work uses S16A Wide-layer data111The S16A data (Wide, Deep, and UltraDeep) will be available in 2019 as a public data release 2. Although Yamashita et al. (2018) used UltraDeep data in addition to Wide data, this work focuses only on Wide data.obtained from 2014 March to 2016 January providing a forced photometry of $g$-, $r$-, $i$-, $z$-, and $y$-band with a 5$\sigma$ limiting magnitude of 26.8, 26.4, 26.4, 25.5, and 24.7, respectively (Aihara et al., 2018b). The HSC–SSP Wide-layer covers six fields (XMM-LSS, GAMA09H, WIDE12H, GAMA15H, HECTOMAP, and VVDS; see Table 1 in Yamashita et al. 2018 for detailed coordinates of each field). The typical seeing is about 0$\arcsec$.6 in the $i$-band and the astrometric uncertainty is about 40 mas in rms. Taking into account the photometric and astrometric flags, Yamashita et al. (2018) eventually extracted 23,795,523 HSC objects in the 154 deg2 for the cross-matching with FIRST (see Section 2.1 in Yamashita et al., 2018, for more detail).

The FIRST project completed radio imaging survey at 1.4 GHz with a spatial resolution of 5$\arcsec$.4 (Becker et al., 1995; White et al., 1997) covering 10,575 deg2 that is completely overlapping with the survey footprint of the HSC-SSP Wide-layer, and the final release catalog of FIRST (Helfand et al., 2015) is publicly available. Before cross-matching with the HSC, Yamashita et al. (2018) made a flux-limited FIRST sample with flux density at 1.4 GHz greater than 1.0 mJy. Taking into account a flag that tells a source is a spurious detection near a bright source, Yamashita et al. (2018) eventually extracted 7072 FIRST objects in the 154 deg2 for the cross-matching with the HSC (see Section 2.2 in Yamashita et al., 2018, for more detail). By cross-matching the HSC S16A Wide-layer catalog and FIRST final data release catalog with a search radius of 1$\arcsec$, 3579 objects (including RGs and radio-loud quasars) were selected (see Section 3 in Yamashita et al., 2018, for more detail).

Before compiling multi-wavelength data, we made a parent RG sample. First, we removed 55 stellar objects (i.e., radio-loud quasars) based on optical morphological information (see Yamashita et al., 2018). For 3579 – 55 = 3524 RGs, we then narrowed down the sample to 2118 objects in three fields with a total area of $\sim 94.7$ deg2 (GAMA09H, WIDE12H, and GAMA15H) where multi-wavelength data are available. We then removed 175 objects that are not covered by FIR observation (see Section 2.1.4), which yielded 1943 RGs. The sky distribution of those 1943 RGs is shown in Figure 2. For those objects, we then complied the multi-wavelength data from $u$-band, near-IR (NIR), MIR, FIR, and radio data, as well as spectroscopic or photometric redshift. After removing 897 objects with photometric data less than 10, and unreliable photometric redshift and/or photometric redshift greater than 1.7 (see Section 2.1.6), we finally selected 1943 - 897 = 1056 objects (hereafter “HSC–FIRST RGs”) with multi-wavelength data and reliable redshift in this work.

2.1.1 $u$-band data

The $u$-band data were taken from the Kilo-Degree Survey (KiDS: de Jong et al., 2013) that is an ESO public survey carried out with the VLT Survey Telescope (VST) and OmegaCAM camera (Kuijken, 2011). We used the Data Release (DR) 3 (de Jong et al., 2017) that consists of 48,736,590 sources with a limiting magnitude of 24.3 mag (5$\sigma$ in a 2″aperture) in $u$-band. The typical full width at half maximum (FWHM) of point spread function (PSF) for $u$-band detected point sources is about 1$\arcsec$222http://kids.strw.leidenuniv.nl/DR3/catalog_table.php.. Before the cross-matching, we extracted 42,252,797 sources with FLAG_U = 0 to ensure clean photometry in $u$-band (see de Jong et al., 2015, 2017, for more detail).

2.1.2 Near-IR data

We compiled NIR data from the VISTA Kilo-degree Infrared Galaxy Survey (VIKING: Arnaboldi et al., 2007) DR3333http://eso.org/rm/api/v1/public/releaseDescriptions/107 that includes 73,747,647 sources in $\sim$1000 deg2 with NIR taken by the VISTA InfraRed Camera (VIRCAM: Dalton et al., 2006). We used $J$-, $H$-, and $K$s-band with a median 10$\sigma$ (Vega) magnitude limit of 20.1, 19.0, and 18.6 mag, respectively. Objects with a PSF FWHM of $<$ 1$\farcs$2 was observed in VIKING. Before the cross-matching, we selected 63,028,265 objects with primary_source = 1 and (jpperrbits $<$ 256 or hpperrbits $<$ 256 or kspperrbits $<$ 256) to ensure clean photometry for uniquely detected objects (see also Toba et al., 2015; Noboriguchi et al., 2019).

2.1.3 Mid-IR data

The MIR data were taken from Wide-field Infrared Survey Explorer (WISE: Wright et al., 2010). We utilized W1 (3.4 $\micron$), W2 (4.6 $\micron$), W3 (12 $\micron$), and W4 (22 $\micron$) data in ALLWISE (Cutri et al., 2014) that consists of 747,634,026 sources. The 5$\sigma$ detection limits444http://wise2.ipac.caltech.edu/docs/release/allwise/expsup/sec2_3a.html in W1, W2, W3, and W4 band are approximately 0.054, 0.071, 0.73 and 5 mJy, respectively. The angular resolutions in W1, W2, W3, and W4 band are 6$\farcs$1, 6$\farcs$4, 6$\farcs$5, and 12$\farcs$0, respectively. We extracted 741,753,366 sources with (w1sat = 0 and w1cc_map = 0) or (w2sat = 0 and w2cc_map = 0) or (w3sat = 0 and w3cc_map = 0), or (w4sat = 0 and w4cc_map = 0) in the AllWISE catalog (Cutri et al., 2014), to have secure photometry at either band (see the Explanatory Supplement to the AllWISE Data Release Products555http://wise2.ipac.caltech.edu/docs/release/allwise/expsup/index.html, for more detail).

2.1.4 Far-IR data

We also used FIR data that were provided by a project of the Herschel Space Observatory (Pilbratt et al., 2010) Astrophysical Terahertz Large Area Survey (H-ATLAS: Eales et al., 2010; Bourne et al., 2016). The data were taken with the Photoconductor Array Camera and Spectrometer (PACS: Poglitsch et al., 2010) at 100 and 160 $\micron$ and with the Spectral and Photometric Imaging REceiver instrument (SPIRE: Griffin et al., 2010) at 250, 350, and 500 $\micron$. The typical PSF FWHMs of 100, 160, 250, 350 and 500 $\micron$ are 11$\farcs$4, 13$\farcs$7, 17$\farcs$8, 24$\farcs$0 ,and 35″.2, respectively. We used H-ATLAS DR1 (Valiante et al., 2016) containing 120,230 sources in the GAMA fields. The 1$\sigma$ noise for source detection (that includes confusion and instrumental noise) is 44, 49, 7.4, 9.4, and 10.2 mJy at 100, 160, 250, 350, and 500 $\micron$, respectively (Valiante et al., 2016).

2.1.5 Ancillary Radio data

The radio data were taken from observations with the Giant Metrewave Radio Telescope (GMRT: Swarup, 1991). We used continuum flux density at 150 MHz ($\sim$1.99 m) provided by the Tata Institute of Fundamental Research (TIFR) GMRT Sky Survey (TGSS) alternative data release (ADR: Intema et al., 2017) that includes 623,604 radio sources in 36,900 deg2. The median rms noise of sources is 3.5 mJy beam*-1* with a spatial resolution of about 25″.

2.1.6 Cross identification of multi-band catalogs

We then cross-identified those catalogs (KiDS, VIKING, ALLWISE, H-ATLAS, and TGSS) with HSC–FIRST RGs666We always use R.A. and Decl. in the HSC catalog as coordinates of HSC–FIRST objects.. By using a search radius of 1$\arcsec$ for KiDS and VIKING, 3$\arcsec$ for ALLWISE, 10$\arcsec$ for H-ATLAS, and 20$\arcsec$ for TGSS, 1051 (54.1%), 1564 (80.5%), 1482 (76.3%), 257 (13.2%), and 471 (24.2%) objects were cross-identified by KiDS, VIKING, ALLWISE, H-ATLAS, and TGSS, respectively. We note that 3/1051 ($\sim$0.3%) and 2/471 ($\sim$0.4%) objects have two candidates of counterpart for VIKING and TGSS sources, respectively within the search radius. We choose the nearest object as a counterpart for such case. For cross-matching with other catalogs (KiDS, ALLWISE, and H-ATLAS), one-to-one identification was realized. The matches by chance coincidence are estimated by generating mock catalogs with random positions, in the same manner as Yamashita et al. (2018). We generated mock catalog of KiDS, VIKING, ALLWISE, H-ATLAS, and TGSS data where source position in each catalog is shifted from the original one to $\pm$1$\arcdeg$ or $\pm$2$\arcdeg$ along the R.A. direction (see Yamashita et al., 2018, for more detail). We then cross-identified HSC-FIRST RGs with those mock catalogs with the exactly same search radii. We found that the chance coincidence of cross-matching with KiDS, VIKING, ALLWISE, H-ATLAS, and TGSS catalog is about 5.0, 1.9, 3.4, 9.3, and 0.6%, respectively.

We also compiled photometric and spectroscopic redshift. For spectroscopic redshift, we utilized the SDSS DR12 (Alam et al., 2015), the Galaxy and Mass Assembly project (GAMA) DR2 (Driver et al., 2011; Liske et al., 2015), and WiggleZ Dark Energy Survey project DR1 (Drinkwater et al., 2010). For photometric redshift, we employed a custom-designed Bayesian photometric redshift code (MIZUKI: Tanaka, 2015) to estimate the photometric redshift (photo-$z$) of HSC–FIRST objects in the same manner as Yamashita et al. (2018) in which we used $z_{\rm best}$ as a photometric redshift (see also Tanaka et al., 2018). In order to perform an accurate SED fitting, we preferentially used spectroscopic redshift. For objects without spectroscopic redshift, we used their $z_{\rm best}$ if they have a reliable photometric redshift, i.e., $0<z_{\rm best}\leq 1.7$777Yamashita et al. (2018) reported that the HSC-SSP photo-$z$ derived by MIZUKI could be secure at $z<1.7$ based on comparison with spectroscopic redshift in COSMOS field (see Section 5.1.2 in Yamashita et al., 2018, for more detail), $\sigma_{z_{\rm best}}/z_{\rm best}\leq 0.1$, and reduced $\chi^{2}$ of $z_{\rm best}$ $\leq$ 5.0. These criteria are optimized based on the comparison with spectroscopic redshift for WERGS sample in Yamashita et al. (2018) (see also Tanaka et al., 2018). However, the influence of the above criteria on physical quantities derived from the SED fitting is still unclear, which will be discussed in Section 4.2.1. In addition to the above redshift (quality) cut, we extracted objects with 3$\sigma$ detection in at least 10 photometric bands among 20 photometric data ($u$, $g$, $r$, $i$, $z$, $y$, $J$, $H$, $K_{\rm s}$-band, and 3.4, 4.6, 12, 22, 100, 160, 250, 350, and 500 $\micron$, and 150 and 1400 MHz) to avoid an overfitting for our SED fitting method (see Section 2.2). Consequently, 1056 HSC–FIRST RGs with multi-band photometry and reliable redshift were left (see Figure 1). Among 1056 objects, the redshifts of 224, 44, and 3 objects were taken from the SDSS DR12, GAMA DR2, and WiggleZ DR1, respectively while the redshifts of the remaining 785 objects were taken from MIZUKI. The HSC-FIRST RG catalog that includes basic information such as redshift and multi-band photometry is accessible through an online service. Format and column descriptions of the catalog are summarized in Table 3.

2.2 SED modeling with CIGALE