Ultraviolet Fe II Emission in Fainter Quasars

 Ultraviolet Fe II emission in fainter quasars: luminosity dependences, and the influence of environments
Monthly Notices of the Royal Astronomical Society (2016), 460, 1428.

Ultraviolet Fe II emission in fainter quasars: luminosity dependences, and the influence of environments

Roger G. Clowes1, Lutz Haberzettl2, Srinivasan Raghunathan3, Gerard M. Williger2,
Sophia M. Mitchell2,4, Ilona K. Söchting5, Matthew J. Graham6, Luis E. Campusano3

1 Jeremiah Horrocks Institute, University of Central Lancashire, Preston PR1 2HE, UK
2Department of Physics and Astronomy, University of Louisville, Louisville, KY 40292, USA
3Observatorio Astronómico Cerro Calán, Departamento de Astronomía, Universidad de Chile, Casilla 36-D, Santiago, Chile
4Department of Aerospace Engineering ACCEND, University of Cincinnati, Cincinnati, OH 45221, USA
5Astrophysics, Denys Wilkinson Building, Keble Road, University of Oxford, Oxford OX1 3RH, UK
6California Institute of Technology, 1200 East California Boulevard, Pasadena, CA 91125, USA

We investigate the strength of ultraviolet Fe II emission in fainter quasars compared with brighter quasars for 1.0 ≤ z ≤ 1.8, using the Sloan Digital Sky Survey (SDSS) DR7QSO catalogue and spectra of Schneider et al. (2010), and the SDSS Faint Quasar Survey (SFQS) catalogue and spectra of Jiang et al. (2006).  We quantify the strength of the UV Fe II emission using the W2400 equivalent width of Weymann et al. (1991), which is defined between two rest-frame continuum windows at 2240–2255 and 2665–2695 Å. The main results are the following. (1) We find that for W2400 ≳ 25 Å there is a universal (i.e. for quasars in general) strengthening of W2400 with decreasing intrinsic luminosity, L3000. (2) In conjunction with previous work by Clowes et al. (2013), we find that there is a further, differential, strengthening of W2400 with decreasing L3000 for those quasars that are members of Large Quasar Groups (LQGs). (3) We find that increasingly strong W2400 tends to be associated with decreasing full width at half maximum (FWHM) of the neighbouring Mg II λ2798 broad emission line. (4) We suggest that the dependence of W2400 on L3000 arises from Lyα fluorescence. (5) We find that stronger W2400 tends to be associated with smaller virial estimates from Shen et al. (2011) of the mass of the central black hole, by a factor of ∼2 between the ultrastrong emitters and the weak. Stronger W2400 emission would correspond to smaller black holes that are still growing. The differential effect for LQG members might then arise from preferentially younger quasars in the LQG environments.


The Fe II problem

The UV Fe II problem in quasars and Seyfert galaxies has been known for more than 30 years – see Wills et al. (1980) and Netzer (1980), and references given in those papers, for some initial discussions of the observational and theoretical aspects. Essentially, the problem is that the UV Fe II emission can vary greatly in strength from one quasar to another, and the variation appears not to simply correspond to iron abundance. Of particular interest are the rare ‘ultrastrong emitters’, a good example of which is the quasar 2226−3905 (Graham, Clowes & Campusano 1996). They represent only ∼6.6 per cent of all quasars in the redshift range 1.0 ≤ z ≤ 1.8 and with i ≤ 19.1 (Clowes et al. 2013). Also of particular interest is the ratio of the UV Fe II flux to that of the nearby Mg II λ2798 doublet, Fe II(UV)/Mg II, because of its potential to allow deduction of the abundance ratio Fe/α, where α refers to the α-elements O, Ne, Mg, etc. The Fe is generally thought to be produced on time-scales ∼1 Gyr by SNe Ia and the α-elements on shorter time-scales by SNe II (e.g. Hamann & Ferland 1999; Hamann et al. 2004). Measurement of the abundance ratio in quasars could then conceivably allow deductions about the time of major star formation relative to the time of quasar activity – an ‘iron-clock’. However, aside from the question of the extent to which the ratio Fe II(UV)/Mg II relates to the abundance ratio (see below), the observations show that it is essentially constant for redshifts z <~ 6.5 and that star formation therefore appears to precede quasar activity by perhaps 0.3 Gyr (e.g. Dietrich et al. 2003; Simon & Hamann 2010; De Rosa et al. 2011).

Note, however, that Reimers et al. (2005) present an abundance analysis for a particular (if unusual) quasar that does not support this notion of an iron-clock. Furthermore, calculations by Matteucci & Recchi (2001) show that the time-scale for the maximum rate of SNe Ia, and hence for the maximum rate of enrichment, depends strongly on the adopted conditions (stellar lifetimes, initial mass function, star formation rate), being ∼40–50 Myr for an instantaneous star- burst, ∼0.3 Gyr for a typical elliptical galaxy, and ∼4–5 Gyr for the disc of a Milky Way (MW)-type galaxy. An instantaneous starburst seems plausible for the central activity of a galaxy that precedes the quasar activity. Matteucci & Recchi (2001) emphasize that the commonly used time-scale of ∼1 Gyr is the time-scale at which the production of iron begins to become important for the solar neighbourhood: it is not the time at which SNe Ia begin to occur. Note that the progenitors of SNe Ia have not yet been established observationally (Nomoto, Kobayashi & Tominaga 2013), although observations can constrain the possibilities for models and abundance yields.

Also, Verner et al. (2009) found evidence that Fe II(UV)/Mg II, for 0.75 < z < 2.2, increases across the interval z ∼ 1.8–2.2, relative to its value at lower redshifts. They interpret this result mainly as a dependence on intrinsic luminosity (ionizing flux) rather than on abundance.

Accounting for the strength of the UV Fe II emission and its variation between quasars is a complicated problem, with many contributing factors such as iron abundance, hydrogen density, hydrogen column density, temperature, ionization flux (excitation flux, continuum shape), microturbulence, Lyα fluorescence, and spatial distribution of the emitting gas (e.g. Verner et al. 2003, 2004; Leighly & Moore 2006; Bruhweiler & Verner 2008; Gaskell 2009; Kollatschny & Zetl 2011, 2013). The quasars with ultra-strong emission, being extreme, may be particularly useful for clarifying the relative importance of different factors. The current view is that probably iron abundance, Lyα fluorescence, and microturbulence are all contributing significantly to the ultrastrong emission.

The most detailed modelling of Fe II emission is by Verner et al. (2003, 2004) and Bruhweiler & Verner (2008). They consider 830 energy levels of the Fe+ ion, corresponding to 344035 transitions. Verner et al. (2004) conclude that iron abundance is not the only factor that can lead to strong emission and, moreover, that it is not even likely to be the dominant factor. Verner et al. (2003) discuss the relative importance of abundance and microturbulence, showing that increasing the iron abundance from solar to five times solar increases the flux ratio Fe II(UV)/Mg II by less than a factor of 2. Conversely, increasing the microturbulence from 5 km s−1 to 25 km s−1 increases the ratio Fe II(UV)/Mg II by more than a factor of 2. [Goad
& Korista (2015) considered the effect of microturbulence on Hβ emission and found that it increased emission across the whole broad line region (BLR), but especially at the smaller radii.] Verner et al. (2003, 2004) suggest that the most reasonable value for the microturbulence is 5–10 km s−1 , while Bruhweiler & Verner (2008) favour 20 km s−1. Ruff et al. (2012) suggest 100 km s−1 as that would produce smooth line profiles. Sigut & Pradhan (2003) and Baldwin et al. (2004) also recognized that abundance is important but unlikely to be dominant.

Early in the history of the Fe II problem, Wills, Netzer & Wills (1985) and Collin-Souffrin & Lasota (1988) concluded that either there was an unusually high abundance of iron or an important mechanism was being overlooked. Microturbulence is one such mechanism since it increases the spread in wavelength of Fe II absorption and thus increases radiative pumping. Earlier than the discussions of microturbulence however, Penston (1987) had proposed that Lyα fluorescence might be the overlooked mechanism. This possibility received observational and theoretical support from Graham et al. (1996) and Sigut & Pradhan (1998), respectively. Lyα fluorescence of Fe II is discussed in detail by Johansson & Jordan (1984) for cool stars and by Hartman & Johansson (2000) for the symbiotic star RR Tel. These papers, and especially Hartman & Johansson (2000), also discuss fluorescence arising from other emission lines, such as C IV λ1548. Bruhweiler & Verner (2008) discuss the significance of the peculiar atomic structure of the Fe+ ion. The 63 lowest energy levels, up to 4.77 eV, are all of even parity, with no permitted transitions between them. These energy levels will be well populated, with the electrons consequently available for pumping to higher levels by the 10.2 eV Lyα line (and also by the continuum).

Johansson & Jordan (1984) and Hartman & Johansson (2000) discuss Lyα fluorescence in connection with stars, but the mechanism is, of course, equally relevant to quasars. However, the width of Lyα is much larger in quasars than in the stars. The most important excitation channels are within ±3 Å of Lyα, as discussed by Johansson & Jordan (1984). For the broad lines of quasars, Sigut & Pradhan (1998) consider the excitation channels within ±50 Å, finding an increase of ∼15 per cent in the ratio Fe II(UV)/Hβ compared with that for ±3 Å (with zero microturbulence in both cases). A particular small emitting region or cloudlet will see the full profile of the Lyα arising from the whole ensemble of cloudlets in the BLR. For quasars, the central concentration of the Lyα flux thus seems likely to be important, to emphasize the important ±3 Å. We might expect that, for quasars, the ratio of FWHM to equivalent width of the Lyα would be a useful central-concentration parameter for associating with the Fe II emission (with low FWHM/EW implying a high central concentration).

In this context of Lyα fluorescence, an interesting case is apparent in Wills et al. (1980). In 0957+561A and 0957+561B the UV Fe II equivalent widths are different by a factor of nearly 2, and the differences are visually obvious in the spectra, although these objects are gravitationally lensed images of the same quasar. This observation suggests that the Fe II emission can vary substantially on time-scales comparable to the ∼1-year time delay between the two images (more precisely, 417 d; Kundić et al. 1997; Shalyapin et al. 2008). The equivalent width of the nearby Mg II λ2798 emission and also of C III] λ1909 appeared unchanged. Iron abundance seems very unlikely to be the cause of the difference. Lyα fluorescence is a possible cause of the difference, given that the Lyα flux in Q0957+561A,B has been observed to be variable on time-scales (observed frame) of weeks (Dolan et al. 2000). The continuum is substantially bluer in the image (A) in which the Fe II is weaker. Wills & Wills (1980) initially attributed this difference in continuum shape to differential reddening along the different light-paths, but, in a note added in proof, subsequently attributed it to the proximity of lensing galaxy G1, at ∼1 arcsec from 0957+561B, in agreement with Young et al. (1980) and then Young et al. (1981). (Note that colour variability has, however, been detected since by Shalyapin, Goicoechea & Gil-Merino 2012.) The lensing of 0957+561 (z = 1.408) arises from a cluster of galaxies (z = 0.355), and G1, the brightest cluster member. G1 is about four times fainter in the R passband than 0957+561B (Walsh, Carswell & Weymann 1979; Young et al. 1980). The spectra of Wills et al. (1980) and Wills & Wills (1980), which have much higher resolution and used smaller apertures than the spectra of Young et al. (1980, 1981), show no indication of the 4000 Å break from G1 affecting
the spectrum of 0957+561B at ∼2250 Å (rest frame), in the vicinity of the low-wavelength end of the UV Fe II feature. The ratio B/A of spectra in figs 1 (single epoch) and 3 (combined epochs) of Wills & Wills (1980) appear consistent with the apparent differences in the Fe II emission not being an artefact arising from G1.

Guerras et al. (2013) have investigated UV Fe II and Fe III emission in the spectra of 14 image-pairs for 13 gravitationally lensed quasars. They find differences for four image pairs, one of which is Q0957+561A,B (apparent for Fe III only). They attribute these differences to gravitational microlensing by stars in the lensing galaxies, but caution that their result depends strongly on one image pair (SDSS J1353+1138A,B). It is not clear why the explanation of the differences has to be microlensing. From statistical considerations, they estimate that the UV Fe II and Fe III emission arises in a region of size ∼4 light-days and suggest that it is located within the accretion disc where the continuum originates (size ∼5–8 light-days). Possible implications of this location for the existence of these ions and for the widths of emission lines are not discussed.

UV Fe II emission and the broad line region

The UV Fe II emission is usually thought to arise in the BLR, but is sometimes attributed instead to an intermediate line region (ILR), between the outer BLR and the inner torus. Graham et al. (1996) and Zhang (2011), for example, favour the ILR interpretation. Photoionization equilibrium implies the temperature of the BLR gas is ∼104 K. For such a temperature the thermal line-widths are ∼10 km s−1 , which is very much smaller than the observed line-widths of ∼1000–20 000 km s−1 . Such a disparity might mean that the BLR contains many (∼108 ) small cloudlets to produce an overall smooth profile (Dietrich et al. 1999), although microturbulence will also cause broadening (Bottorff et al. 2000), as will Rayleigh and Thomson scattering (Gaskell & Goosmann 2013). In reality, the distribution of the gas is likely to be fractal (Bottorff & Ferland 2001).

The view of the structure of the BLR has changed somewhat over the years, in substantial part because of the results of reverberation mapping of Seyfert galaxies (e.g. Peterson 2006). In the old view there is a roughly spherical distribution of the cloudlets, each with stratification of the ionization. Although much is uncertain, in the modern view there is a more flattened distribution of cloudlets (Gaskell 2009; Kollatschny & Zetl 2011, 2013), with a general stratification of the ionization and density, such that high-ionization and high density correspond to smaller distances from the central BH and low-ionization and low density correspond to larger distances (Gaskell 2009). The flattening is more pronounced for low-ionization lines (Kollatschny & Zetl 2013). The BLR could be a thick disc of cloudlets or could be bowl-shaped (Goad, Korista & Ruff 2012; Pancoast et al. 2012, 2014). In the RPC model (for radiation pressure confinement; Baskin, Laor & Stern 2014) the cloudlets are replaced by a BLR that is a single stratified slab. High-ionization lines tend to be broader than low-ionization lines in the same quasar. The low-ionization optical Fe II and Mg II emission is thought to arise in the same outer regions (Gaskell 2009; Korista & Goad 2004; Cackett et al. 2015). There is some evidence from reverberation mapping of Seyfert galaxies that the UV Fe II arises at smaller radii in the BLR than the optical Fe II and is correspondingly broader (e.g. Vestergaard & Peterson 2005; Barth et al. 2013). The flattened distribution rotates about the central BH, but there is also (macro)turbulent motion of ∼1000 km s−1 or more (Kollatschny & Zetl 2013). Turbulence (macroturbulence) refers to bulk motion measured orthogonal to the plane of rotation, but will be present in the plane of rotation too (Gaskell 2009; Kollatschny & Zetl 2013). Microturbulence refers to (non-thermal) motion within cloudlets or between adjacent cloudlets (Bottorff et al. 2000). The rotation velocity exceeds the turbulence velocity by factors of a few times. Failure to account for the turbulence will lead to the mass of the BH being overestimated (Kollatschny & Zetl 2013). In the case of a BLR viewed face-on (i.e. viewing along the rotation axis, sometimes also expressed as ‘pole-on’), one is seeing mainly the turbulent motion in the emission lines. The FWHM that we measure for the broad emission lines is thus a function of viewing angle (Wills & Browne 1986).

The amount of gas in the BLR exceeds by a large factor (∼103 – 104 ) that needed to account for the line emission. The notion of ‘locally optimally emitting cloud’ (LOC) (Baldwin et al. 1995) is that the regions that are emitting strongly are those where the conditions to do so are optimal. A change in the continuum luminosity can then result in an apparent change of the scale-size of the BLR, simply because another region is then the optimal emitter. From simple modelling of photoionization, the scale-size should increase approximately as the square root of the ionizing luminosity (Peterson 2006). A particular cloudlet or small region will receive the Doppler-shifted Lyα profile of the whole BLR ensemble (i.e. broadened by the macroturbulent and rotational velocities) but sliced by narrow absorption lines from intervening cloudlets (Gaskell, private communication). The cloudlet itself will be emitting with a line-width comparable to the sound velocity or microturbulence velocity.

The investigation of metallicity in the nuclear regions of quasars proceeds by the analysis of the broad emission lines and intrinsic (i.e. associated with the quasar, not intervening) narrow absorption lines (e.g. Hamann & Ferland 1999; Hamann et al. 2004). The growth of the central supermassive BH is believed to arise from accretion following large-scale events in the galaxies – mergers and interactions – and from quieter, secular processes such as flows along bars. Infall of (possibly pristine) gas might also be involved. The gas that approaches the nucleus either forms stars or it settles into the dusty torus, and from there spirals into the accretion disc and the BH. The distinction between the torus and the accretion disc is that the temperature of the accretion disc exceeds the sublimation temperature of the dust. The BLR is actually the turbulent gas above the accretion disc, and which rotates with it. (Boundaries between accretion disc, torus and BLR are related to the cooling process that is dominant: continuum-cooling for the accretion disc; thermal emission from dust for the torus; line-cooling for the BLR.) Some small fraction of the incoming gas is expelled in a wind that removes angular momentum and allows the accretion. Dynamical models of the BLR suggest inflows, outflows, and orbital motion (e.g. Pancoast et al. 2012, 2014; Grier et al. 2015). The metallicity of the BLR should be that of the torus. Czerny & Hryniewicz (2011) discuss, and give a good illustration in their fig. 1, the accretion disc extending from the BH to the outer edge of the torus, where the torus dust sublimates, with the clouds (low-ionization clouds at least) of the BLR ‘boiling’ (inflowing and outflowing) from its outer parts. In their interpretation the atmosphere of the outer parts of the disc is cool enough for dust to be present. The outflow of the boiling is driven locally by the accretion disc – a dust-driven wind – but the driving force is cut-off at large heights by the central radiation sublimating the dust, leading to infall. Outflow and inflow would lead to turbulence. Note, however, that for NGC 4151, Schnülle et al. (2013) find evidence that the inner torus is actually located beyond the sublimation radius and that it does not have a sharp boundary (see also Kishimoto et al. 2013).

Hamann et al. (2004) consider the bright phase of a quasar to correspond to a final stage of accretion during which the BH roughly doubles its mass. This phase lasts for ∼6 × 107 yr for accretion at ∼50 per cent of the Eddington rate. The metallicities of the BLR and the torus can be expected to be typical of the central parts of galaxies at the redshifts concerned. The more massive galaxies can be expected to have higher central metallicities because of repeated episodes of star formation, and because the deeper potential well increases the retention of gas for recycling. The observations indicate nuclear metallicities of solar to several times solar. It is also possible that mergers and interactions could drive relatively unprocessed gas into the nuclear regions, which would reduce the metallicity. Radiation pressure could, in principle, concentrate metals relative to hydrogen in some regions (e.g. in BAL outflows or in the outer parts of accretion discs), but observationally the effect is not currently known to occur (Baskin 2012; Baskin & Laor 2012). Quasar activity is believed to be preceded by star-formation activity. One might then expect to see a dependence on redshift of the metallicity of quasars, but none has so far been detected for z <~6.5. Enrichment of the BLR appears to be completed well before (0.3–1 Gyr) the visible quasar activity (e.g. Dietrich et al. 2003; Simon & Hamann 2010; De Rosa et al. 2011). There is no indication of further enrichment from continuing star formation.

Current results

log_l_3000_w2400_f1_cFigure 1. A plot of W2400 against log10L3000 for the sample DR7QSO10. L3000 is the intrinsic continuum flux, in units of erg s-1 , at 3000 Å in the rest frame. The shading indicates the kernel-smoothed densities of points in a 64 × 64 grid, because the number of points, 25700, is too high for an ordinary scatterplot to be a useful illustration. Linear contours for the densities are also shown. The plot has been restricted to 44.6 ≤ log10L3000 ≤ 47.0 and 0 ≤ W2400 ≤ 70 Å for clarity. Note that the highest Fe II emitters tend to favour lower values of log10L3000. The plot also shows, as points (crosses), W2400 against log10L3000, for the sample SFQS 10 (80 quasars). The same trend is apparent for the SFQS 10 quasars.

log_l_3000_f1Figure 2. The distributions of log10L3000 for the 1531 ultrastrong W2400 emitters of the DR 7QSO10 sample (solid blue histogram) and for the 7086 strong W2400 emitters (hatched red histogram). L3000 is the intrinsic continuum flux, in units of erg s-1 , at 3000 Å in the rest frame  Both are density histograms. The bin size is 0.1 in the logarithm. The histograms have been restricted to 45.0 ≤ log10L3000 ≤ 47.0 Å for clarity. The distribution for the ultrastrong emitters is clearly shifted to lower values of log10L3000 relative to that for the strong emitters.

fwhm2798_w2400_f1_cFigure 3. A plot of W2400 against fwhm2798, the FWHM of the Mg II λ2798, emission for the sample DR7QSO10. The shading indicates the kernel-smoothed densities of points in a 64 × 64 grid, because the number of points, 25700, is too high for an ordinary scatterplot to be a useful illustration. The plot has been restricted to 0 ≤ fwhm2798 ≤ 140 Å and 0 ≤ W2400 ≤ 70 Å for clarity. Note that the highest Fe II emitters tend to favour relatively narrow Mg II emission lines.

log10mbh_w2400_f1_cFigure 4. A plot of W2400 against log10(MBH/Msun), where MBH is the adopted “fiducial” virial BH mass from Shen et al. (2011), for the sample DR7QSO10. The shading indicates the kernel-smoothed densities of points in a 64 × 64 grid, because the number of points, 25670, is too high for an ordinary scatterplot to be a useful illustration. The plot has been restricted to 7.5 ≤ log10(MBH/Msun) ≤ 10.5 and 0 ≤ W2400 ≤ 70 Å for clarity. Note that increasing W2400 emission seems to be associated with decreasing values of MBH .


We might expect narrow emission lines to have some correspondence with relatively less massive central BHs – either to AGN that intrinsically have low BH masses or to younger quasars that are still growing their central BHs and have not yet reached their mature state (e.g. Mathur 2000). In Fig. 9, we plot W2400 against log10(MBH/Msun) for DR7QSO10, where MBH is the adopted “fiducial” virial BH mass from Shen et al. (2011). MBH is actually determined by a FWHM, acting as a proxy for virial velocity, and by a continuum luminosity, acting as a proxy for BLR radius. Again, the number of points (25670 of the previous 25700 have a measurement of MBH) is too high for an ordinary scatterplot to be a useful illustration and the plot shows instead the kernel-smoothed densities of points in a 64 × 64 grid. From Shen et al. typical errors in the BH masses, propagated from the measurement uncertainties in the continuum luminosities and the FWHMs, are ∼0.05–0.2 dex, but the additional statistical uncertainty in the calibration of virial masses is ∼0.3–0.4 dex. Nevertheless, there does appear to be a trend in Fig. 9, with average ultrastrong emitters having smaller BH masses than average weak emitters by a factor of ∼2. Clearly, the ultrastrong emitters are not in the low-BH-mass category (108Msun) of AGN, and we might instead expect that they are still growing their BHs. In that case, increasing W2400 emission would have an association with increasingly young quasars. We further suggest that this possible interpretation in terms of younger quasars helps to clarify the nature of LQGs. LQGs – regions with an overdensity of quasars – would then be interpreted quite naturally as regions that contain a higher proportion of young quasars than field regions.


The conclusions of this paper are the following.

(i) We find that for W2400 ≳ 25 Å there is a universal (i.e. for quasars in general) strengthening of W2400 with decreasing intrinsic luminosity, L3000.

(ii) In conjunction with the work presented by Clowes et al. (2013, we find that there is a further, differential, strengthening of W2400 with decreasing L3000 for those quasars that are members of Large Quasar Groups, or LQGs.

(iii) We find that increasingly strong W2400 tends to be associated with decreasing FWHM of the neighbouring Mg II λ2798 broad emission line.

(iv) On the assumption that this trend that W2400 increases as the FWHM of Mg II decreases is true also for the FWHM of the Lyα emission line, we suggest that this dependence of W2400 on intrinsic luminosity arises from Lyα fluorescence. We make this suggestion because: the wavelength region 1216 ± 3 Å is important for Lyα fluorescence of the Fe II emission; for a given Lyα equivalent width, narrower Lyα emission would enhance the flux in this region, and hence fluorescence of the Fe II , relative to broader Lyα emission.

(v) We find that stronger W2400 tends to be associated with smaller virial estimates of the mass of the central BH, by a factor of ∼2 between the ultrastrong emitters (smaller BHs) and the weak. The effect for W2400 can then be associated with a range of masses for the central BHs. Stronger W2400 emission would correspond to smaller BHs that are still growing. They will be in fainter quasars, with relatively narrow broad emission lines, leading to the enhanced Lyα fluorescence of the UV Fe II emission. The differential effect for LQGs might then arise from a different mass distribution of the BHs, corresponding, plausibly, to an age distribution that emphasises younger quasars in the LQG environments.


Baldwin J., Ferland G., Korista K., Verner D., 1995, ApJ, 455, L119
Baldwin J. A., Ferland G. J., Korista K. T., Hamann F., LaCluyzé A., 2004, ApJ, 615, 610
Barth A. J. et al., 2013, ApJ, 769, 128
Baskin A., 2012, in Chartas G., Hamann F., Leighly K. M., eds, ASP Conf. Ser. Vol. 460. AGN Winds in Charleston. Astron. Soc. Pac., San Francisco, p. 159
Baskin A., Laor A., 2012, MNRAS, 426, 1144
Baskin A., Laor A., Stern J., 2014, MNRAS, 438, 604
Bottorff M., Ferland G., 2001, ApJ, 549, 118
Bottorff M., Ferland G., Baldwin J., Korista K., 2000, ApJ, 542, 644
Bruhweiler F., Verner E., 2008, ApJ, 675, 83
Cackett E. M., Gültekin K., Bentz M. C., Fausnaugh M. M., Peterson B. M., Troyer J., Vestergaard M., 2015, ApJ, 810, 86
Clowes R. G., Raghunathan S., Söchting I. K., Graham M. J., Campusano L. E., 2013, MNRAS, 433, 2467
Collin-Souffrin S., Lasota J.-P., 1988, PASP, 100, 1041
Czerny B., Hryniewicz K., 2011, A&A, 525, L8
De Rosa G., Decarli R., Walter F., Fan X., Jiang L., Kurk J., Pasquali A., Rix H. W., 2011, ApJ, 739, 56
Dietrich M., Wagner S. J., Courvoisier T. J.-L., Bock H., North P., 1999, A&A, 351, 31
Dietrich M., Hamann F., Appenzeller I., Vestergaard M., 2003, ApJ, 596, 817
Dolan J. F., Michalitsianos A. G., Nguyen Q. T., Hill R. J., 2000, ApJ, 539, 111
Gaskell C. M., 2009, New Astron. Rev., 53, 140
Gaskell C. M., Goosmann R. W., 2013, ApJ, 769, 30
Goad M. R., Korista K. T., 2015, MNRAS, 453, 3662
Goad M. R., Korista K. T., Ruff A. J., 2012, MNRAS, 426, 3086
Graham M. J., Clowes R. G., Campusano L. E., 1996, MNRAS, 279, 1349
Grier C. J. et al., 2015, ApJ, 806, 111
Guerras E., Mediavilla E., Jimenez-Vicente J., Kochanek C. S., Muñoz J. A., Falco E., Motta V., Rojas K., 2013, ApJ, 778, 123
Hamann F., Ferland G., 1999, ARA&A, 37, 487
Hamann F., Dietrich M., Sabra B. M., Warner C., 2004, in McWilliam A., Rauch M., eds, Origin and Evolution of the Elements, Carnegie Observatories Astrophysics Series Volume 4. CUP, Cambridge, p. 440
Hartman H., Johansson S., 2000, A&A, 359, 627
Jiang L. et al., 2006, AJ, 131, 2788
Johansson S., Jordan C., 1984, MNRAS, 210, 239
Kishimoto M. et al., 2013, ApJ, 775, L36
Kollatschny W., Zetl M., 2011, Nature, 470, 366
Kollatschny W., Zetl M., 2013, A&A, 558, A26
Korista K. T., Goad M. R., 2004, ApJ, 606, 749
Kundić T. et al., 1997, ApJ, 482, 75
Leighly K. M., Moore J. R., 2006, ApJ, 644, 748
Mathur S., 2000, MNRAS, 314, L17
Matteucci F., Recchi S., 2001, ApJ, 558, 351
Netzer H., 1980, ApJ, 236, 406
Nomoto K., Kobayashi C., Tominaga N., 2013, ARA&A, 51, 457
Pancoast A. et al., 2012, ApJ, 754, 49
Pancoast A., Brewer B. J., Treu T., Park D., Barth A. J., Bentz M. C., Woo J.-H., 2014, MNRAS, 445, 3073 (with erratum, 2015, MNRAS, 448, 3070)
Penston M. V., 1987, MNRAS, 229, 1p
Peterson B. M., 2006, in Alloin D., Johnson R., Lira P., eds, Physics of Active Galactic Nuclei at All Scales. Springer, Berlin, p. 77
Reimers D., Janknecht E., Fechner C., Agafonova I. I., Levshakov S. A., Lopez S., 2005, A&A, 435, 17
Ruff A. J., Floyd D. J. E., Webster R. L., Korista K. T., Landt H., 2012, ApJ, 754, 18
Schneider D. P. et al., 2010, AJ, 139, 2360
Schnülle K., Pott J.-U., Rix H.-W., Decarli R., Peterson B. M., Vacca W., 2013, A&A, 557, L13
Shalyapin V. N., Goicoechea L. J., Koptelova E., Ullán A., Gil-Merino R., 2008, A&A, 492, 401
Shalyapin V. N., Goicoechea L. J., Gil-Merino R., 2012, A&A, 540, A132
Shen Y. et al., 2011, ApJS, 194, 45
Sigut T. A. A., Pradhan A. K., 1998, ApJ, 499, L139
Sigut T. A. A., Pradhan A. K., 2003, ApJS, 145, 15
Simon L. E., Hamann F., 2010, MNRAS, 407, 1826
Verner E., Bruhweiler F., Verner D., Johansson S., Gull T., 2003, ApJ, 592, L59
Verner E., Bruhweiler F., Verner D., Johansson S., Kallman T., Gull T., 2004, ApJ, 611, 780
Verner E., Bruhweiler F., Johansson S., Peterson B., 2009, Phys. Scr. T 134, 014006
Vestergaard M., Peterson B. M., 2005, ApJ, 625, 688
Walsh D., Carswell R. F., Weymann R. J., 1979, Nature, 279, 381
Weymann R. J., Morris S. L., Foltz C. B., Hewett P. C., 1991, ApJ, 373, 23
Wills B. J., Browne I. W. A., 1986, ApJ, 302, 56
Wills B. J., Wills D., 1980, ApJ, 238, 1
Wills B. J., Netzer H., Uomoto A. K., Wills D., 1980, ApJ, 237, 319
Wills B. J., Netzer H., Wills D., 1985, ApJ, 288, 94
Young P., Gunn J. E., Kristian J., Oke J. B., Westphal J. A., 1980, ApJ, 241, 507
Young P., Gunn J. E., Kristian J., Oke J. B., Westphal J. A., 1981, ApJ, 244, 736
Zhang X.-G., 2011, ApJ, 741, 104

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s