Radio Dichotomy in Quasars with H$\beta$ FWHM greater than $15,000$ km s^-1

Since the very first discovery of quasars (Schmidt, 1963), the radio bi-modality of the quasar population has been discussed in the literature (e.g., White et al., 2000; Brotherton et al., 2001; Lacy et al., 2001; Cirasuolo et al., 2003a, b; Doi et al., 2013; Tadhunter, 2016; Hardcastle & Croston, 2020). While the principal difference between the radio-loud (RL) and radio-quiet (RQ) quasars lies in the strength of their radio jets (e.g., Bridle et al., 1994; Mullin et al., 2008), several studies have been undertaken to address the source of this dichotomy (e.g., Kellermann et al., 1989; Miller et al., 1990; White et al., 2000; Lacy et al., 2001; Ivezić et al., 2002; White et al., 2007; Laor & Behar, 2008; Zamfir et al., 2008; Baloković et al., 2012; Zhang et al., 2021). The dichotomy has been studied in the context of black hole (BH) mass (e.g., McLure & Jarvis, 2004), accretion rate (e.g., Sikora et al., 2007; Hamilton, 2010), host galaxy properties (e.g., Sikora et al., 2007; Lagos et al., 2009; Kimball et al., 2011), as well as black hole (BH) spin (e.g., Blandford & Znajek, 1977; Blandford, 1990; Garofalo et al., 2010). It has been proposed that there may be multiple mechanisms responsible for powering the radio emission, for example it may come from a combination of small-scale jets as well as star formation regions (Cirasuolo et al., 2003a; Sabater et al., 2019).

In the cosmological context, the role of galaxy mergers and host dark matter halos in the RL-RQ dichotomy has also been discussed (e.g., Hughes & Blandford, 2003; Shen et al., 2007; Hickox et al., 2009; Lagos et al., 2009; Garofalo et al., 2010; Hamilton, 2010). Using host galaxy photometry and optical spectroscopy of the nuclear region, several recent works revealed a consistent mass excess of the central BHs powering RL quasars compared to RQ quasars (e.g., Dunlop et al., 2003; Chiaberge et al., 2005; Jarvis & McLure, 2006). Other studies have reached similar conclusions (e.g., Magorrian et al., 1998; Laor, 2000b; McLure & Dunlop, 2001) regarding the mass distributions of RL and RQ quasars.

Some authors have shown that BH mass is correlated with radio luminosity (e.g., Franceschini et al., 1998; McLure et al., 1999; Laor, 2000a; Lacy et al., 2001; McLure & Jarvis, 2004) while some studies report the contrary (e.g., Ho, 2002; Oshlack et al., 2002; Woo & Urry, 2002; Snellen et al., 2003). In the context of this dichotomy, there have been proposals of radio-intermediate sources, which essentially implies the distribution of radio loudness to be quasi-continuous (Falcke et al., 1996; Brotherton et al., 2001; Lacy et al., 2001; Gürkan et al., 2019). One of the key reasons for which it has been particularly difficult to establish the bimodality, is the absence of a universally accepted definition of radio loudness. A ratio of the radio to optical luminosity has been used most often while the absolute value of the radio luminosity has been assumed to be the characteristic parameter in some other cases. However, for both of those parameters the RL-RQ boundary has not been universal over various studies in the literature (e.g., Baloković et al., 2012).

Radio emission in quasars is shown to be related with the optical emission lines (e.g., Osterbrock, 1977). In radio galaxies, a correlation between emission line luminosities and radio jet luminosities has been established (e.g., Baum & Heckman, 1989; Saunders et al., 1989; Rawlings & Saunders, 1991; Willott et al., 1999; Zirbel & Baum, 1995; Buttiglione et al., 2010; Sikora et al., 2013). Therefore, emission line diagnostic has been proposed by some authors as a distinct method to define RL quasars. However, studies reveal that there can be significant scatter in the radiative vs kinetic emission correlations (see e.g., Rawlings & Saunders, 1991; Punsly & Zhang, 2011; Mingo et al., 2014; Mingo et al., 2016; Gürkan et al., 2015). It has been shown that jets do produce emissions beyond the radio, and are capable of photoionizing the surrounding interstellar medium, but through shock-ionization they can also produce emission lines (e.g., Mingo et al., 2011).

Zamfir et al. (2008) showed that RL quasars are located at a specific region on an H$\beta$ full width half maximum (FWHM) versus $R_{Fe}$ plane. We followed Zamfir et al. (2008) to define $R_{Fe}=F_{Fe}/F_{H\beta}$, where $F_{H\beta}$ is the optical continuum flux at 5100 Åand $F_{Fe}$ is the FeII flux at 4861 Å , calculated using the optical continuum flux at 5100 Å  and H$\beta$ slope. Among the sample they used roughly 78% of RL quasars have H$\beta$ FWHM $>$ 4000 km s^-1. While some RQ quasars are located in that part of the plane, the restricted location of RL quasars is statistically significant.

Accreting BH systems have been found to reside preferentially on a plane in a space comprising disc luminosity, jet power and BH mass (e.g., Maccarone et al., 2003; Merloni et al., 2003; Terashima & Wilson, 2003). This is known as the fundamental plane of BH activity. It shows that the disc activity, jet emission, and BH mass are connected with each other. However, the exact nature of the above connection is not well understood. In general relativistic magnetohydrodynamic simulations of jet launching and collimation (e.g., Komissarov, 2007; McKinney et al., 2012) it is revealed that the conversion of accreted material into collimated jets may depend on multiple variables or a combination thereof, including, BH mass (e.g., Dunlop et al., 2003; Chiaberge & Marconi, 2011), BH spin (e.g., Blandford & Znajek, 1977; Fernandes et al., 2011), and disc activity (e.g., Lin et al., 2009; Fernandes et al., 2011). During the accretion process, gas in the disc gets heated and strong ultraviolet (UV) to optical continuum emission is radiated (Pringle, 1981; Malkan & Sargent, 1982; Ma & Wang, 2013). Optical spectroscopic observations of the continuum and line emission of RL and RQ quasars have shed light on the connection between the accretion disc and jet of these systems (Kawakatu et al., 2009). It has been shown that the radio loud fraction (RLF, hereafter) of quasars decreases with decreasing optical luminosity (Jiang et al., 2007).

In the preliminary work of Chakraborty & Bhattacharjee (2021), we used a sample of RL and RQ quasars selected through the Sloan Digital Sky Survey (SDSS) and performed a systematic study of their optical continuum and line emission to search for clues in understanding the dichotomy in their radio emission. In particular, we studied the BH properties of RL and RQ objects selected through their H$\beta$  FWHM. Our results showed that the RLF increases with FWHM, and hence to further understand this result we select a sub-sample of extreme objects in the catalog with H$\beta$ FWHM greater than 15,000 km s^-1 and study the bolometric luminosity, optical continuum luminosity and fiducial virial BH mass (Vestergaard & Peterson, 2006) distributions of these extreme objects. Gas in the broad line region (BLR) of quasars is photoionized by the emission from the central continuum source at ultraviolet frequencies, and gives rise to prominent recombination lines observed in the quasar spectra (Sulentic et al., 2000). Due to their proximity to the central SMBH the BLR gas has high speed ($10^{3}-10^{4}$ km s^-1) which results in the Doppler broadening of the emission lines. Thus our sample consists of quasars with large values of the speed of the BLR gas.

In the current work, we investigate whether the radio luminosity, bolometric luminosity, BH mass, and radio loudness parameter of the RL quasars in our sample of high broad line (HBL hereafter; with FWHM greater than $15,000$ km s^-1) quasars are related to each other while comparing them with their relevant RQ counterparts. We construct the composite SDSS spectra of the quasars in our sample and search for any difference between the RL and RQ populations in terms of their mean spectrum. Finally, we compare our HBL sample with the rest of the H$\beta$ broad emission line objects (non-HBL hereafter) in terms of BH mass, Eddington ratio, bolometric luminosity, H$\beta$ FWHM, and $R_{Fe}$ to search for possible clues of higher RLF.

Throughout this paper we use $\Omega_{\Lambda}$ = 0.7, $\Omega_{0}$ = 0.3 and h = 0.7 as the cosmological parameters (Spergel et al., 2007). In §2, we give a brief description of our sample and present our results in §3. In §4 we discuss the results and summarize our conclusions.

In this section we discuss our data sets and sample selection.

Here we present our results for the HBL as well as the entire H$\beta$ broad emission line sample of Shen et al. (2011) for RL and RQ quasars.

In Figure 1, we show that the radio loud fraction increases with the FWHM of broad H$\beta$ emission line for z $<0.75$ quasars identified in the SDSS. To further investigate this effect, we selected a subsample of extreme objects (FWHM $>$ 15,000 km $s^{-1}$) to study their properties in detail. For our sample, number of RL quasars is 56 among a total of 97, which implies an RLF of 57%, much larger than that of the larger sample of all SDSS quasars matched with FIRST.On the other hand, the BH mass of the quasars in our subsample is, on average, higher than that of the non-HBL SDSS quasars while the optical continuum luminosities are lower on average. Consequently, the Eddington ratio of the quasars in our subsample $\sim 10^{-2}-10^{-1}$, is significantly lower than the average of that of the non-HBL quasars. Thus, for equivalent BH mass range, quasars in our sample tend to have lower luminosity and hence a lower accretion rate.

It has been observed in stellar mass BH accreting systems that the X-ray spectrum becomes harder when the luminosity goes below a threshold, the value of which is in the range $0.01-0.1$ $L_{edd}$, where $L_{edd}$ is the Eddington luminosity. It has been thoroughly discussed in the literature that this hard state may be caused by a switch of the nature of the accretion flow from the standard geometrically thin optically thick radiatively efficient disk to a geometrically thick optically thin advection-dominated accretion flow (ADAF) (Narayan & Yi, 1994, 1995; Abramowicz et al., 1995; Kato et al., 1997; Narayan, 2005). In the ADAF, unlike a standard disk, heat produced in the accretion through viscous dissipation is not emitted. Instead the energy is confined to the accreted material and moves toward the BH along with the flow. As a result the accretion luminosity is low (Narayan & Yi, 1995; Nakamura et al., 1997; Narayan, 2005). ADAF has been used to explain observations of various classes of AGN as well. ADAF has the additional property that the accreted material is not strongly bound to the BH. Hence, there is a higher probability of producing disk wind and jet in AGN containing an ADAF (Meier, 2001; Narayan, 2005). Although our findings show some characteristics of ADAF, we note that ADAF models are unlikely to produce the strong optical emission observed in quasars. Studies of jet systems led to magnetically arrested accretion (e.g., Tchekhovskoy et al., 2011). We refer the reader to Hardcastle & Croston (2020) for discussion on accretion processes in accreting systems with radio jets. We propose that our subsample of quasars with high FWHM of emission lines may contain magnetically arrested accrection flows, which is consistent with having a very large fraction being radio-loud.

In our subsample of high broad line quasars, there is no significant difference in mass among the two classes, namely, radio-loud and radio-quiet. However, the RL quasars have larger bolometric luminosity than the RQ ones as computed by Shen et al. (2011). Furthermore, we computed the combined spectrum of SDSS quasars with FWHM $>$ 15,000 km $s^{-1}$. It shows the mean optical continuum luminosity to be higher for RL quasars, in comparison to RQ quasars. Given that the difference in the bolometric or optical continuum luminosity is mainly due to that of the accretion disk, the above implies some connection between a stronger jet and a brighter disk. Accretion disc-jet connection in AGN has often been studied using X-ray and radio observations (e.g., Naik & Rao, 2000; Marscher et al., 2002; Chatterjee et al., 2009, 2011; Chatterjee et al., 2018). In AGN, X-rays are not produced in the disk itself but in a distribution of energetic particles located near the disk termed corona, which is assumed to be energetically and dynamically connected to the disk. On the other hand, optical emission is directly produced in the disk and hence may be better suitable for probing its connection to the jet. Inoue et al. (2017) have found a strong correlation between the disc and the jet, using bolometric luminosity (as inferred from the optical luminosity of the disc) and 1.4 GHz radio luminosity as an indicator of the jet emission. We note that our findings are consistent with this result although our sample size is not very large.

We found that the emission line properties of the composite spectra of the RL and RQ quasars of our HBL sample are similar. The major difference between the two classes has been observed in the strength of the [O III] narrow line. The composite spectrum of the HBL RL quasars exhibits a stronger [O III] line compared to that of the RQ ones. This is consistent with previous findings that RL quasars have a larger narrow line region and stronger narrow emission lines than their RQ counterparts (e.g., Heckman et al., 1991; Brotherton et al., 2001). This may be caused by the interaction of the radio jet with the narrow line gas, as inferred from observations indicating an alignment between the radio jet axis and the narrow emission line region (e.g., Cimatti et al., 1993). On the other hand, Schulze et al. (2017) have shown an enhancement in [O III] line strength by a factor of at least 1.5 in a radio-loud compared to a radio-quiet sample, both of which have similar redshifts, black hole mass, and accretion rate as indicated by the optical continuum luminosity. They conclude that the enhanced [O III] line strength is due to higher extreme UV luminosity at 35 eV for the RL objects. They suggest this is caused by the radio-loud quasar population having systematically higher black hole spin and consequently their radiative efficiency being larger, consistent with the paradigm (Wilson & Colbert, 1995; Sikora et al., 2007; Tchekhovskoy et al., 2010) that the radio loud objects have higher black hole spin.

In this work we employ a sample of high broad line quasars and search for hints in the observed higher radio loud fraction in these objects compared to the general population of quasars. Our study suggests that a magentically arrested flow like accretion and a stronger connection between the disc and the jet may result in this higher radio function. We propose that this might provide a clue to the broader question of understanding radio dichotomy in the quasar population.

AC thanks Jaya Maithil from the University of Wyoming, Swamtrupta Panda from the Laboratório Nacional de Astrofísica - MCTIC and Preeti Kharb of National Centre for Radio Astrophysics, for their valuable inputs. The authors thank the referee for several important feedback which greatly helped in improving the draft.

SC acknowledges support from the Department of Science and Technology through the SERB-ECR grant, SERB-CRG grant and from the Department of Atomic Energy for the BRNS grant. RC and SC are grateful to the Inter University Center for Astronomy and Astrophysics (IUCAA) for providing infrastructural and financial support along with local hospitality through the IUCAA-Associateship program. RC thanks Presidency University for support under the Faculty Research and Professional Development (FRPDF) Grant, and ISRO for support under the AstroSat archival data utilization program. RC acknowledges financial support from BRNS through a project grant (sanction no: 57/14/10/2019-BRNS) and thanks the project coordinator Pratik Majumdar for support regarding the BRNS project.

The data underlying this article will be shared on reasonable request to the corresponding author.