The quasar SDSS J142507.32+323137.4 : dual AGNs?
aa r X i v : . [ a s t r o - ph . C O ] N ov Research in Astron. Astrophys.
Vol. x No. XX , 000–000 R esearchin A stronomyand A strophysics The quasar SDSS J142507.32+323137.4 : dual AGNs?
Zhixin Peng , , Yanmei Chen , , Qiusheng Gu , and Chen Hu Department of Astronomy, Nanjing University, Nanjing 210093, China; [email protected] Key Laboratory of Modern Astronomy and Astrophysics (Nanjing University), Ministry ofEducation, Nanjing 210093, China Key Laboratory for Particle Astrophysics, Institute of High Energy Physics, Chinese Academy ofSciences, 19B Yuquan Road, Beijing 100049, China
Received [year] [month] [day]; accepted [year] [month] [day]
Abstract
We analyze the optical spectrum of type 1 QSO SDSS J1425+3231. This ob-ject is interesting since its narrow emission lines such as [O
III ] λλ ∼ times broader. The separation between the blue and redcomponents is ∼ km/s with blue component ∼ times broader than the red one. TheH β emission can be separated into four components: two for the double-peaked narrowline and two for the broad line which comes from the broad line region (BLRs). The blackhole mass estimated from the broad H β emission line using the typical reverberation map-ping relation is . × M ⊙ , which is consistent with that derived from parameters of[O III ] λ β emission line is mainly contributed by the primary black hole (traced bythe blue component) while the broad H β component of the secondary black hole (tracedby the red component) is hard to be separated out considering a resolution of ∼ Key words: galaxies: active — galaxies: individual (SDSS J142507.32+323137.4) —quasars: emission lines
Early in 1980s, the double-peaked narrow emission lines in active galactic nuclei (AGNs) have beenreported by Heckman et al. (1981, 1984), and they suggested biconic outflows as an origin of doublepeaked AGNs. Greene & Ho (2005) pointed out that there are about 1% local AGNs have double-peakednarrow emission lines. Since narrow emission lines are generally believed to be produced by clouds inthe narrow line regions (NLRs), they suggested that ∼
1% AGNs have disk-like NLRs as a simplestexplanation. The ∼
1% double-peaked AGN fraction have also been found by Zhou et al. (2004) andWang et al. (2009). Especially, the flux ratios of the two peaks of [O
III ] λ Z.X. Peng, Y.M. Chen & Q.S. Gu et al. 2010) since on the one hand, they are unavoidable results of hierarchical cosmology model whichhas gained great success; on the other hand, with the launch of X-ray and infrared space astronomicalfacilities (Chandra, Spitzer, and Herschel Space telescopes), it is now possible for us to resolve the dualAGNs with kpc-scale separation. Observational evidence for dual AGNs includes: spatially resolvedsystems in which both supermassive black holes (SMBHs) can be identified directly and spatially unre-solved systems in which the dual AGN model can explain various phenomena (see Komossa 2006 fora detail review). Most recently, Colpi & Dotti 2009 further summarize the observations and numericalsimulations of dual and binary black holes. So far, a few unambiguous cases have been found, such asNGC 6240 (Komossa et al. 2003), J0402+379 (Rodriguez et al. 2006, 2009; Morganti et al. 2009),EGSD2 J142033+525917 (Gerke et al. 2007), EGSD2 J141550+520929 (Comerford et al. 2009a),COSMOS J100043+020637 (Comerford et al. 2009b), and other four dual AGNs (see Liu et al. 2010).There are other interesting sources, which need more certification by future observations. Zhou et al.(2004) connected both SDSS and VLBA data, suggesting SDSS J1048+0055 is a dual AGN system anddouble-peaked narrow emission lines could be an effective way of selecting dual AGN candidates. Xu& Komossa (2009) analyze the line structures and flux ratios of SDSS J1316+1753 in detail, discussingall the possible origins of the double peaks. Furthermore, SDSS J1536+0441 is the only source in whichtwo broad line systems have been found, and it is suggested to be a binary black hole system whichis separated by 0.1pc with a orbital period of 100 yr (Boroson & Lauer 2009). However it is also themost controversial case, which brings about a vast deal of debate (Chornock et al. 2009, 2010; Wrobel& laor 2009; Decarli et al. 2009; Lauer & Boroson 2009; Tang & Grindlay 2009; Gaskell 2010; Dotti &Ruszkowski 2010; Bondi & P´erez-Torres 2010).The peculiar emission-line spectrum of SDSS J1425+3231 is noticed in the course of searching fordual AGN candidates in the SDSS QSO sample. SDSS J1425+3231 shows all its strong narrow emissionlines with double-peaked, it is hard to make conclusions on the structures of weak emission lines due tothe signal-to-noise ratio (S/N). We analyze the spectra of SDSS J1425+3231 in §
2. The black hole massis estimated in §
3. The possible origins of the double-peaked line profiles of this source are discussed in §
4. The results are summarized in §
5. Throughout this paper, a cosmology with H = 70 km s − Mpc − , Ω M = 0.3, and Ω Λ = 0.7 is adopted. SDSS J1425+3231 is a broad line QSO with SDSS pipeline redshift of z = 0 . , all its strong narrowemission lines show double-peaked profiles. In this section, we describe the procedure of spectral fitting.The steps of our analysis are as follows: (1) the spectrum is corrected for foreground Galactic extinctionand shifted to the rest-frame by using z = 0 . ; (2) the continuum of the spectrum is modeled by threecomponents (Hu et al. 2008) and subtracted, the aim is to separate the contribution of continuum andemission line spectrum; (3) multiple Gaussian components are used to fit the emission lines. Step (2)and (3) are described in more detail below. The continuum is modeled as F λ = F PL λ ( F , α ) + F BaC λ ( F BE , τ BE ) + F Fe λ ( F Fe , FWHM Fe , V Fe ) . (1)where F PL λ = F ( λ ) α is a featureless power law, F is the flux at 5100 ˚A and α is the spectralindex. The second and third terms represent the Balmer continuum and Fe emission, respectively.For wavelength shortward of the Balmer edge λ < λ BE = 3646 ˚A, the Balmer continuum can beexpressed as F BaC λ = F BE B λ ( T e )(1 − e − τ λ ) (Grandi 1982; Dietrich et al. 2002). F BE is a normalizationcoefficient for the flux at λ BE , B λ ( T e ) is the Planck function at an electron temperature T e , τ λ = τ BE ( λλ BE ) is the optical depth at wavelength λ , τ BE is the optical depth at the Balmer edge. T e isassumed to be T e = 15 , K. The two free parameters in the Balmer continuum are F BE and τ BE .At λ > λ BE , blended higher-order Balmer lines give a smooth rise from ∼ DSS J142507.32+323137.4 3 (Wills et al. 1985) in the spectrum. However, our fitting windows do not include this region, actually ourresults are not influenced by the higher order Balmer lines.The optical and ultraviolet Fe II template ( F IZw1 λ ) from NLS1 I ZW 1 is used to subtract the Fe IIemission from the spectra (Boroson & Green 1992; Vestergaard & Wilkes 2001). The I ZW 1 templateis broadened by convolving with a Gaussian function G : F Fe λ = F IZw1 λ ∗ G ( F conv , FWHM conv , V conv ) , (2)where F conv , FWHM conv , and V conv are the flux, width and peak velocity shift of the Gaussian function.The parameters of Fe in equation (1) can be expressed as follows: the flux of the Fe emission, F Fe , it isthe multiplication of F conv and the flux of the template; the shift of the Fe spectrum, V Fe = V conv , andthe FWHM of the Fe lines FWHM Fe = p FWHM + FWHM .In total, there are seven parameters in the continuum model, they are fitted by minimizing χ .The fitting windows include: 2470–2625, 2675–2755, 2855–3010, 3625–3645, 4170–4260, 4430–4650,5080–5550, and 6050–6200 ˚A. These windows are free of strong contaminant lines. Figure 1 showsthe result of the continuum decomposition. The Galactic extinction and redshift corrected spectrum isshown in the top panel. The spectrum in the fitting window is plotted in green. The three componentsof the continuum are shown in blue. The best fit model is shown in red. The middle panel shows theresidual spectrum, namely the pure emission-line spectrum. We will analyze it in the next step. In thebottom panel, we subtract the power law and the Balmer continuum, zoom in Fe-only spectrum in thewavelength range 4100–5600 ˚A. The Fe model is shown in red. We measured the H β and [O III ] λ III ] λ σ and flux. Each [O III ] λ III ] λ III ] λ ∼ km s − in velocity space, and the σ of blue component is417 km s − , about 2.2 times broader than the red component. The broad wing (orange) peaks at roughlythe same position as that of the blue component. The σ of the wing is 628 km s − . We use four-Gaussiancomponents to model the asymmetric profile of the H β emission, two for the blue and red narrow peaks,they are forced to have the same width as the corresponding [O III ] λ III ], we do not fit a wing to the narrow H β component sinceon the one hand, comparing with the broad H β , the strength of the wing can be neglected; on the otherhand, the resolution of the SDSS spectrum is not high enough to separate such a weak component fromthe broad H β . The broad H β are fitted with the two other Gaussian components, see the pink lines. Thefitting parameters of each Gaussian component is shown in table 1. In this section, we use three different methods to estimate the mass of the central black hole.The first method we used is based on the virial theorem and R BLR − L relation which is cal-ibrated with the reverberation mapping data. The BLR radius R BLR is estimated from the continuumluminosity at 5100 ˚A ( L = 1 . × erg s − ˚ A − ) using the R BLR − L relation given byBentz et al. (2006), then the black hole mass is estimated from M BH = f R BLR ∆ V G , where f is thescaling factor which is introduced to characterize the unclear kinematics and geometries of the BLRs.Actually, the value of f changes with the shape of the line in use, f = 3 . is a mean value suggested Z.X. Peng, Y.M. Chen & Q.S. Gu
Fig. 1
SDSS spectrum of SDSS J1425+3231 plotted as observed flux versus rest wavelength.The top panel shows the Galactic extinction and redshift corrected spectrum. The spectrum inthe fitting window is plotted in green. The three components of the continuum are shown inblue while The best fit model is shown in red. The middle panel shows the residual spectrum,namely the pure emission-line spectrum. The bottom panel shows the spectrum after subtractingthe power law and the Balmer continuum in the wavelength range 4100–5600 ˚A. The Fe modelis shown in red. The flux density F λ is given in units of 10 − erg s − cm − ˚A − by Collin et al. (2006). ∆ V = 1325km s − is the second moment of the BLR H β profile, which isreconstructed from the two pink components in Figure 2. This method gives a value of . × M ⊙ .The M BH − L relation (see Eq. 9 of Peterson et al. 2004) predicts a mass of . × M ⊙ .We also use velocity dispersion of gas in the NLR, namely σ of [O III ] λ σ ∗ (Nelson 2000), to estimate the central black hole mass based on the famous M BH − σ ∗ relation which is in the form of log( M BH /M ⊙ ) = 8 .
13 + 4 . σ ∗ / (Tremaine et al. DSS J142507.32+323137.4 5
Fig. 2
Emission-line fitting of SDSS J1425+3231. For [O
III ] λ β are fitted with four Gaussian components, two forthe double-peaked narrow line (red and blue) and another two for the broad line (see the two pinkcomponents). The best fit model is shown in green.2002). The blue component gives a mass of . × M ⊙ with σ = 177 .
40 km s − while the redcomponent indicates a mass of . × M ⊙ with σ = 80 .
18 km s − . This possibility is very unlikely for two reasons. First, the red component of [O
III ] λ . × erg s − , while the σ of this line indicates a black hole mass of . × M ⊙ . If thiscomponent comes from a background AGN, its accretion rate should be extremely super-Eddington,which is hard to explain (see section 4.3 for the explanation under the dual AGN scenario). Second,Dotti & Ruszkowski (2010) examined the superposition model of double-peaked emission line AGNsbased on galaxy clusters from the Millennium Run, finding that the fraction of superimposed galaxypairs peaks at about z = 0 . and decreases rapidly since z = 0 . . Considering the redshift of SDSSJ1425+3231, ∼ . , the possibility of superposition should be very low. Z.X. Peng, Y.M. Chen & Q.S. Gu
Table 1
Emission line properties of SDSS J1425+3231
Property H β [O III ] [O
III ]4861.33 4958.91 5006.84Blue systemLine center ( ˚A) 4861.77 ± ± ± km s − ) 417.17 ± ± ± a ± ± ± ± ± ± km s − ) 188.21 ± ± ± a ± ± ± ± ± km s − ) - 1281.42 ± ± a - 150.33 ± ± ± km s − ) 1545.02 ± a ± b Line center ( ˚A) 4861.08 ± km s − ) 4463.45 ± a ± Notes: a In units of erg s − . b Note only the H β need the very broad component in appearance. Another possible explanation for the double-peaked narrow emission could be special NLR geometriessuch as biconical outflows and disk-like NLRs. In such hypothesis, there is only one AGN to illuminatethe NLR gas which is moving toward and away from us, forming the blue and red components in theobserved narrow lines.Certain nearby Seyfert and star forming galaxies are known to have biconic outflow induced double-peaked emission lines. The examples are found not only from the spatial resolved spectra which takesalong the minor axis of a galaxy (e.g. Cecil 1988; Cecil et al. 1990; Veilleux et al. 1994, 2001; Colbertet al. 1996) but also from the spectrum of the whole galaxy (e.g. Duric & Seaquist 1988; Axon et al.1998). In the scenario of biconic outflows, we would expect that the blue and red components have thesame velocity dispersion since they are illuminated by the same AGN. However, in SDSS J1425+3231,the blue component is three times broader than the red component, which is conflicted with the outflowmodel. On the other hand, the outflow studies find that the NLRs are stratified strongly in ionization andvelocity so that high-ionization lines, such as [O
III ] are originated near the AGN with higher velocityand low-ionization lines, such as H β and [O II ], are originated further from the AGN with lower velocity(e.g., Komossa et al. 2008). In SDSS J1425+3231, we find that [O III ] and H β are consistency in velocityoffsets within error bars, and apparently no ionization stratification is observed as expected in AGNdriven outflows. In addition, the [O III ] λ . × erg s − for the blue system and . × erg s − for the red system. These are typical values for the emission from the whole NLRsof bright Seyfert galaxies and quasars, but it is hard to image that the whole NLRs are outflowing.The observed double-peaked narrow emission lines can be accounted for in a disk-like NLR model.The rotating disk model predicts that the blue and red components have similar width, which is conflictwith the current data. Furthermore, the red and blue components are expected to be (almost) equallyshifted with respect to the true cosmological galaxy redshift in this scenario. We are lack of host galaxyinformation since it is over-shined by the central AGN, however under the disk-like NLR model, the DSS J142507.32+323137.4 7 broad line should be at the redshift of the galaxy, in between the blue and red components, and not soclose to the the blue component as we have seen from the data.
The final picture we want to suggest for SDSS J1425+3231 is that this is a dual AGN system. Theprimary black hole in this system has a mass of ∼ M ⊙ , the observed blue narrow component in[O III ] λλ β component and the wing of [O III ] λλ β component and σ of the blue component supports this idea. We have not observed the broad emissionlines from the BLR of the secondary black hole ( ∼ M ⊙ ), the possible reasons for this includes: (1)the S/N of SDSS spectra is not high enough for us to separate the broad emission line of the secondaryblack hole from primary one; (2) the secondary black hole is a type 2 AGN in which the BLR is obscuredby the dusty torus.The only issue about the dual AGN system is that, for the secondary black hole, comparing with the[O III ] λ M ⊙ is a little bit low. This indicates the secondary black hole isaccreting in a super Eddington regime. At first sight, this is conflicted with the normal accretion theory.However, we should note that in the dual AGN case, the separation between the two black holes is inkpc scale, the NLR gas of the secondary black hole can be affected by the primary black hole, namelythe primary black hole can illuminate the NLR of the secondary black hole. Type 1 QSO SDSS J1425+3231 has double-peaked narrow emission lines. In this paper, we analyzethe SDSS spectrum of this object, discussing the origins of its double-peaked line structure. We argueagainst the possibility of superposition of two objects, biconic outflow and disk like NLR, suggestingthat this is a dual AGN system.In this system, the primary black hole has a mass of ∼ M ⊙ , the observed broad lines belong toit. The secondary black hole is much smaller, its mass is in the oder of M ⊙ . The secondary blackhole could be a type 2 AGN whose BLR is obscured, or we failed to separated the broad emission line ofthe secondary black hole from that of the primary black hole due to the resolution of the SDSS spectra.In the future observation, high spatial resolution two-dimensional optical spectrum, and imagingin the optical, radio, X-ray would help us figure out whether SDSS J1425+3231 contains dual AGNs.Moreover, the ultimate confirmation or rejection of the dual AGN interpretation, which predict the vari-ations in the line profiles, will come from multi-epoch spectroscopic monitoring. Acknowledgements
We thank the anonymous referee for suggestions that led to improvements in thispaper. The research is supported by the National Natural Science Foundation of China (NSFC) underNSFC-10878010, 11003007 and 10633040, and the National Basic Research Program (973 programNo. 2007CB815405). This research has made use of NASA’s Astrophysics Data System BibliographicServices and the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet PropulsionLaboratory, California Institute of Technology, under contract with the National Aeronautics and SpaceAdministration. This work is based on observations made with the
Spitzer Space Telescope
Z.X. Peng, Y.M. Chen & Q.S. Gu for Advanced Study, the Japan Participation Group, The Johns Hopkins University, Los AlamosNational Laboratory, the Max- Planck-Institute for Astronomy (MPIA), the Max-Planck-Institute forAstrophysics (MPA), New Mexico State University, University of Pittsburgh, Princeton University, theUnited States Naval Observatory and the University of Washington.