In our extreme luminosity Hamburg-ESO sample (Marziani et al., 2009; Sulentic et al., 2017), we have identified four cases of highly radiating quasars that show an extreme behavior, i.e., a high Eddington ratio and a strong blue asymmetry [c(1/2) <–4,000 km s⁻¹; centroid at half intensity] in the Civλ1549 profile (Sulentic et al., 2017). The most extreme case corresponds to the quasars HE0359-3959, with z = 1.5209, log(L_bol) = 47.6 erg s⁻¹ and a R_FeII = 1.12. It is cataloged as an A3 source (see Figure 1).

The aim of this paper is to analyze the spectral behavior of an extreme xA source, the quasar HE0359–3959. We performed multicomponent fits in a wide spectral range: UV, optical and Near–Infrared (Section 2); which gives us information about the dynamics and the physical conditions of the broad line region (BLR) (Section 3). In Section 4, we summarize the main results of our work.

Ultraviolet (UV), optical, and Near–Infrared spectra were observed with the Very Large Telescope (VLT-ESO). Optical and Near–Infrared spectra were obtained with the Infrared Spectrometer And Array Camera (ISAAC; decommissioned in 2013) using a slit of 0.6″. The near–infrared spectrum was observed in 2010 in the K band with a total exposure time of 1,120 s. The optical spectrum was observed in 2004 in the J band with a total exposure time of 3,600 s. For the ultraviolet spectrum we used the Focal Reducer and low dispersion Spectrograph (FORS1) and a slit of 1.0″ with a total exposure time of 1,440 s. It was observed in 2008. The data reduction was done using the iraf package. The procedures followed are explained in Marziani et al. (2009), Martínez-Aldama et al. (2015), and Sulentic et al. (2017).

We perform multicomponent fits using specfit, an iraf routine (Kriss, 1994) to get the information of the most important emission lines. In each spectral range we fit a local continuum. The FWHM of all the broad components (BC) for Hβ, Aliiiλ1860, Siiii]λ1892, Civλ1549, and Siivλ1397 was taken equal. In the Figure 2, we present the multicomponents fits after continuum subtraction, for the Civλ1549 and Ca ii triplet range. The rest of the fits will be shown in an upcoming paper.

Low–ionization lines (LIL) have an ionization potential (IP) ≲20 eV. The Hβ line is the prototype of LIL. In population A3 and A4 sources Hβ has associated a blueshifted component (Bachev et al., 2004). In the case of HE0359-3959, the blueshifted component has a contribution to the total flux of ~9%, and shows a centroid a half intensity of c(1/2)≈−500 ± 70 km s⁻¹.

The Feii (IP~16 eV) has an important contribution in the optical and near–infrared regions. To reproduce it we used the templates modeled by Marziani et al. (2009) and Garcia-Rissmann et al. (2012) for the optical and near–infrared ranges, respectively. Several works have found (Persson, 1988; Ferland and Persson, 1989; Joly, 1989; Dultzin-Hacyan et al., 1999; Martínez-Aldama et al., 2015) a close relationship between the Feii and the NIR Ca ii λ8498, λ 8542, and λ8662 Å triplet. This relation is very well appreciable in this object: as well as the optical Feii is strong, the NIR Ca ii triplet also is. It is the first time where we observe the Ca ii triplet lines isolated at high redshift. Strong intensities of both ions imply an extremely low-ionization degree (U < 10 ⁻²; U: ionization parameter) and a high density (n_H ~ 10^11–13 cm⁻³) (Baldwin et al., 2004; Matsuoka et al., 2007; Martínez-Aldama et al., 2015).

In the UV region, the 1900Å blend is formed by two intermediate–ionization lines (IIL; IP ~ 20–40 eV), Aliiiλ1860 and Siiii]λ1892, which are accompanied by Ciii]λ1909 and some Feiii transitions. In this blend we appreciate a blueshifted component. This component should be most likely associated with Aliiiλ1860. Respect to Aliiiλ1860, the blueshifted component has a contribution of the total profile of 60%. The centroid a half intensity is c(1/2)≈−3,200 ± 250 km s⁻¹, which indicates the presence of an outflow generated by radiation forces presented in the intermediate–ionization lines (Marziani et al., 2017). On the other hand, considering the high intensity of Aliiiλ1860, Siiii]λ1892, Ca ii and Feii, it could suggest a possible chemical enrichment of the BLR (Juarez et al., 2009).

High ionization lines (HIL; IP > 40 eV), Civλ1549, Heiiλ1640, and Siivλ1397, show a prominent blueshifted component. We find that the blue component has a contribution of 76, 62, and 57% to the total flux of Civλ1549, Heiiλ1640, and Siivλ1397, respectively. The Civλ1549 reaches c(1/2) ~−6,000 ± 500 km s⁻¹, while Heiiλ1640 and Siivλ1397 c(1/2) ~−4,000 ± 550 km s⁻¹. The velocities reached are ones of the highest found in the literature (Richards et al., 2011; Coatman et al., 2016; Sulentic et al., 2017). Then, it indicates that the full profile is dominated by an outflow and suggests the disk plus wind scenario (Gaskell, 1982; Richards et al., 2002, 2011).

In order to study the physical properties of the quasar HE0359-3959, we built a grid of photoionization simulations using the CLOUDY code (Ferland et al., 1998, 2013). For our simulations we considerer a Mattews and Ferland continuum (Mathews and Ferland, 1987), a plane-parallel geometry, a metallicity 5Z⊙ with an overabundance of Al and Si with respect to carbon (by a factor of three), and a column density of N_c = 10²³ cm⁻². See Negrete et al. (2012) for more details. Our simulations span the density range 7.00 ≤ log (n_H) ≤ 14.00 and −4.5 ≤ log (U) ≤ 0.00 for the ionization parameter, in intervals of 0.25 dex. More details about the cloudy simulations can be found in Negrete et al. (2014). Using the UV lines, we define three groups of diagnostic ratios::

In Figure 3 is shown the result of the simulations. We obtained that the flux ratios are intersected in log(n_H) = 12.32 cm⁻³ and log(U) = −2.95. Compared to not highly radiating AGNs (Negrete et al., 2013), this source shows a high density and a low ionization parameter, which marks a different behavior in the BLR, probably causing by the slim disk hosted in these kind of sources. Taking into account the high intensity of Aliiiλ1860, Feii, and Ca ii we conclude that effectively the low–ionization emitter zone has a high density and low–ionization parameter.

Negrete et al. (2012) proposed a new method to determine the size of the BLR (r_BLR) and the black hole mass (M_BH) based on the product n_H·U and independently of redshift. This method gives similar results to the obtained from the classical methods such as reverberation mapping at low–z (Negrete et al., 2014). Knowing the product of n_H·U obtained from the cloudy simulations, we compute the size of the BLR (r_BLR) and considering the FWHM of the broad components as the velocity dispersion, we can get the black hole mass (M_BH) and the Eddington ratio. The size of the BLR is log(r_BLR) = 18.37 ± 0.04 cm and the black hole mass is log(M_BH) = 9.52 ± 0.41 M_⊙. These values are in agreement with the ones found for a large xA sample at high-redshift (Martínez-Aldama et al., in preparation).

The Eddington ratio for this source is L/L_Edd = 0.74 ± 0.11. Considering that it shows a c(1/2)~−6,000 ± 500 km s⁻¹ for Civλ1549, we confirm the directly proportional relation between c(1/2) and L/L_Edd. Indicating that L/L_Edd could be the driver of the outflows (Sulentic et al., 2017).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.