Stefano Bianchi AGN without BLR New clues for
Stefano Bianchi AGN without BLR New clues for a new scenario A. Marinucci, G. Matt, F. Nicastro, F. Panessa, X. Barcons, F. Carrera, F. La Franca, N. Sacchi, A. Corral, L. Monaco, A. Ruiz, M. Brightman, K. Nandra, A. Wolter “The X-ray Universe 2011” Berlin – June 28 th 2011
The detection of broad optical lines in the polarized spectrum of the archetypical Seyfert 2, NGC 1068, led Antonucci (1993) to postulate the equivalence between Type 1 and Type 2 AGN, any observational difference due to obscuration along the line-of-sight to the source However, about half of the brightest Seyfert 2 galaxies appear not to have hidden broad-line regions (HBLR) in their optical spectra, even when high-quality spectro-polarimetric data are analysed (e. g. Tran 2001, 2003)
Are NHBLR 'different' from HBLR sources? Many pieces of evidence have been presented to support the view that HBLR Sy 2 s are intrinsically more powerful than non-HBLR Sy 2 s, rather than 'less obscured' or 'starburstcontaminated' (see e. g. Tran 2001, 2003) Nicastro (2000) It appears that HBLRs share many similar large-scale characteristics with Sy 1 s, being their counterparts as predicted by the UM, while non-HBLRs could be a different class of sources Tran (2003) In a model proposed by Nicastro (2000), the creation of the BLR is connected with disk instabilities occurring in proximity of a transition radius at which the accretion disk changes from gas-pressure dominated to radiation-pressure dominated Since this transition radius becomes smaller than the innermost stable orbit for very low accretion rates (and therefore luminosities), very weak AGN should lack the BLR
The existence of a critical threshold in luminosity and/or Eddington rate has been confirmed on observational grounds (e. g. Nicastro 2003, Bian & Gu 2007, Shu et al. 2007, Wu et al. 2011) However, the separation between the two classes is still controversial, and results depend on the choice of the sample and the methods adopted to estimate the accretion rates Nicastro et al. 2003 Indeed, most of these works suffer from highly uncertain estimates of the Eddington ratios, due to very indirect methods to estimate the bolometric luminosity and/or BH masses (e. g. from the [OIII] emission lines)
et al. (2001) Marinucci et al. in prep. Our sample is based on the spectro-polarimetric surveys performed by. Tran (1992, 1995, 2001, 2003), Young et al. (1996), Moran et al (2000) and Lumsden In order to have reliable estimates of the accretion rate for our objects, we require: üa measure of the stellar velocity dispersion σ* to estimate the BH mass ügood SNR (>150 counts) X-ray observations from the most sensitive instruments to estimate the bolometric luminosity Following these requirements, our final sample is composed by 21 sources with polarized hidden broad lines and 18 sources without
Marinucci et al. in prep. When using only Compton-thin sources, a separation between HBLR and non-HBLR sources is clear, both in luminosity (KS test: 99. 5%) and in accretion rate (KS test: 99. 9%) The largest separations between the two distributions are at log. Lbol=43. 90 and log Lbol/LEdd=− 1. 88 Marinucci et al. in prep. Including the Compton-thick sources (correction factor 70), the separation between HBLR and non. HBLR is less sharp (KS tests: 98. 2% and 99. 1%) We note that no HBLR source falls below the Eddington ratio threshold , while there is a number of non-HBLR Compton-thick sources with Eddington ratios larger than the threshold
This mixing could be due to the larger uncertainties involved in determining the intrinsic luminosity of Compton-thick sources, because of the unknown geometry of the absorbers and reflectors Marinucci et al. in prep. This is indeed what happens if the bolometric luminosities are estimated from [OIII] However, these uncertainties could explain the presence of few non-HBLR closely above threshold, but it is unlikely that the much higher accretion rates are wrongly estimated This is also supported by the fact that no HBLR is found below the threshold, which is what we would expect if the mixing were caused by random uncertainties
Are non-HBLR Compton-thick sources 'more obscured'? By comparing the average Lx/L[OIII] in the Comptonthick sources of our sample to that derived by Lamastra et al. (2010) for Comptonthin sources, we estimate a correction factor of ~70 for the intrinsic luminosity of these objects Non-HBLR Compton-thick sources preferentially have lower values of the Lx/L[OIII] ratio. However, a K-S test on the two distributions is not significant We also verified if only the 9 non-HBLR Compton-thick sources that fall in the region with Lbol/LEdd>-1. 88 could be different than all the other Compton-thick sources in our sample with respect to this ratio, but a K-S test is again not significant
It appears that there are two classes of non-HBLR: those with low accretion rates, really lacking the BLR, those with high accretion rate, likely hosting the BLR, but something prevents us from observing it Heisler et al. 1997 The second class is composed only by Compton-thick sources, so the nucleus is severely obscured by intervening absorbers. This could be explained within the framework of standard UM: more inclined sources should intercept a larger column density of the torus, and may obscure the medium responsible for the scattering of the BLR photons (e. g. Shu 2007). This scenario would also explain the (albeit not significantly) lower Lx/L[OIII] ratios found in non. HBLR Compton-thick sources, as a geometrical effect due to the shrinking of the reflecting area with respect to the lineof-sight
If the BLR cannot exist in weakly accreting AGN, we would expect genuine Seyfert 2 galaxies, without any evidence of obscuration of their nuclei 'Unabsorbed Seyfert 2 s' do exist, and the best examples (where the lack of the optical broad lines and of the X-ray obscuration are unambiguosly found in simultaneous observations with high SNR) have Eddington rates lower than the accretion rate threshold estimated by Marinucci et al. (in prep. ) NGC 3147 log(Lbol/LEdd)~-4 Bianchi et al. (2008) Simultaneous observations: it's not an artefact of variability!
Matt et al. (in prep. ) NGC 3147 is NOT Compton-thick: Neutral iron Kα EW: ~130 e. V X-ray/IR/[OIII] flux diagrams No excess above 10 ke. V Strong iron line: outer disk, torus (without BLR!)? Adapted from Panessa & Bassani (2002)
Q 2131− 427 Panessa et al. 2010 log(Lbol/LEdd)~-2. 6 NGC 3660 log(Lbol/Ledd)~--2 (see also Brightman & Nandra 2008) No iron lines NGC 3660: EW<100 e. V Q 2130: unconstrained Bianchi et al. in prep.
How many? Among the 'good' candidates Panessa & Bassani (2002) sample: 17 Sy 2 s galaxies with N H<1022 cm− 2 and very 'unlikely' to be Compton-thick Excluding the sources where a column density (even if much lower than the one expected from the optical properties) is actually measured and misclassified sources 5 sources: only 1 is a true unabsorbed (NGC 3147) The others are either Compton-thick or misclassified (starburst) 'Naked' AGN from Hawkins (2004): Sy 2 s with rapid optical variability 6 sources: 1 unabsorbed Sy 2 (Q 2130 -427), 3 still to be properly verified Brightman & Nandra (2008) sample: Sy 2 s from the IRAS 12 μm sample of Rush, Malkan & Spinoglio (1993) with good quality X-ray data, and N H < 1022 cm− 2), 6 sources: 2 unabsorbed Sy 2 s (NGC 3660 and again NGC 3147)
How many more? Apparently, not so many. . . Among all the X-ray unobscured AGN observed in targeted observations by XMM-Newton (CAIXA: Bianchi et al. 2009), only 6 have an Eddington rate lower than the accretion rate threshold found by Marinucci et al. (in prep. ) NGC 3147 & NGC 3660: unabsorbed Sy 2 s! NGC 4579 & NGC 5033: unabsorbed Sy 2 candidates in the Panessa & Bassani (2002) sample. However, weak broad optical lines are reported for both objects (see e. g. Shi 2010): very interesting! PG 1011 -040: barely below the threshold, it is strongly X-ray weak with respect to its multi-wavelength emission, so that the accretion rate derived from its X-ray luminosity is likely to be orders of magnitude lower than the real one (Gallagher et al. 2001, Vasudevan & Fabian 2007) NGC 7213: the only known object where the iron Kα line is entirely produced in the BLR, which is visible in the optical spectrum ( Bianchi 2008)! However, this is also the only bright Sy 1 known to lack reprocessing features from a Compton-thick torus (Bianchi et al. 2003, 2004 – Lobban et al 2010)
Summary It appears that the presence of the BLR depends on the accretion rate: lowly accreting sources cannot form the BLR However, the detectability of the BLR in polarized light does also depends on geometry/obscuration Two classes of non-HBLR: those with low accretion rates, really lacking the BLR, those with high accretion rate, likely hosting the BLR, but something prevents us from observing it Unabsorbed Sy 2 s exist and are the counterparts of 'real' non-HBLR They do have low accretion rate They are a small fraction of the Seyfert population (? )
- Slides: 15