Does the X-ray outflow quasar PDS 456 have a UV outflow at 0.3c?
MMNRAS , 1–11 (2015) Preprint 18 January 2018 Compiled using MNRAS L A TEX style file v3.0
Does the X-ray outflow quasar PDS 456 have a UV outflowat 0.3c?
Fred Hamann , (cid:63) George Chartas , James Reeves and Emanuele Nardini Department of Physics & Astronomy, University of California, Riverside, CA 92507, USA Department of Physics & Astronomy, College of Charleston, Charleston, SC 29424, USA Astrophysics Group, School of Physical and Geographical Sciences, Keele University, Keele, Staffordshire, ST5 5BG, UK INAF – Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5, I-50125 Firenze, Italy
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
The quasar PDS 456 (at redshift ∼ log N H ( cm − ) > , and large kinetic energies that could be importantfor feedback to the host galaxy. A UV spectrum of PDS 456 obtained with the HubbleSpace Telescope in 2000 contains one well-measured broad absorption line (BAL) at ∼ α at v ≈ . c or N v λ ≈ . c .However, we use photoionisation models and comparisons to other outflow quasars toshow that these BAL identifications are problematic because other lines that shouldaccompany them are not detected. We argue that the UV BAL is probably C iv λ ≈ . c . This would be the fastest UV outflow ever reported, but its speedis similar to the X-ray outflow and its appearance overall is similar to relativisticUV BALs observed in other quasars. The C iv BAL identification is also supportedindirectly by the tentative detection of another broad C iv line at v ≈ . c . Thehigh speeds suggest that the UV outflow originates with the X-ray UFO crudely 20to 30 r g from the central black hole. We speculate that the C iv BAL might form indense clumps embedded in the X-ray UFO, requiring density enhancements of only (cid:38) iv BAL might therefore be the first detection of low-ionisation clumpsproposed previously to boost the opacities in UFOs for radiative driving.
Key words: line: formation – quasars: individual: PDS 456 – quasars: absorptionlines – quasars: general
Accretion disk outflows are an important part of the quasarphenomenon that might drive “feedback” to regulate blackhole growth and host galaxy evolution (e.g., Di Matteo et al.2005; Hopkins et al. 2008; Hopkins & Elvis 2010; Debuhret al. 2012; Rupke & Veilleux 2013). The outflows are of-ten studied in the rest-frame UV via blueshifted broad ab-sorption lines (BALs) or their narrower cousins called “mini-BALs”(with a nominal boundary near full width at half min-imum FWHM ∼ km s − , Weymann et al. 1991; Koristaet al. 1993; Crenshaw et al. 2003; Hamann & Sabra 2004;Trump et al. 2006; Knigge et al. 2008; Gibson et al. 2009, andrefs. therein). These features appear most often at moderatevelocity shifts v < . c , but relativistic BALs and mini-BALs (cid:63) E-mail: [email protected] at v ∼ . –0.2 c have been measured in a small but growingnumber of quasars (Hamann et al. 1997a, 2013; Rodr´ıguezHidalgo 2008; Rodr´ıguez Hidalgo et al. 2011; Rogerson et al.2016).The basic physical properties for these outflows can bedifficult to determine due to limited wavelength coverageand line saturation that is masked by partial line-of-sightcovering of the background light source(s). The strongestmeasured UV lines are typically C iv λ v λ vi vi absorption tends to be as strong orstronger than C iv (Section 3.4 below, also Baskin et al. Throughout this paper we treat unresolved doublets as singlelines, such that, for example, C iv λ iv λ © a r X i v : . [ a s t r o - ph . GA ] J a n F. Hamann et al. viii λ
774 and Mg x λ
615 havebeen observed in a few cases with suitable spectral coverage(e.g., Hamann et al. 1997b; Telfer et al. 1998; Arav et al.2001). Measurements of the low-abundance line P v λ log N H ( cm − ) (cid:38) . ,across a wide range observed BAL strengths (Hamann 1998;Leighly et al. 2009, 2011; Borguet et al. 2012; Capellupoet al. 2014, 2017; Moravec et al. 2017, but see also Arav etal. 2001).X-ray observations have revealed another variety ofultra-fast outflows (UFOs) that reach relativistic speeds inluminous quasars (Chartas et al. 2002, 2009; Reeves et al.2009; Gofford et al. 2013, 2015; Tombesi et al. 2010, 2013,and refs. therein). UFOs are also challenging to study be-cause they are highly variable and highly-ionised to the pointwhere the only strong absorption features appear at X-raywavelengths. They are often characterized by Fe K-shell ab-sorption features with derived total column densities in therange log N H ( cm − ) ∼ to 24 (see refs. above). UFOs alsoappear to have generally very large kinetic energies, suffi-cient to drive feedback effects in the quasar host galaxies(Tombesi et al. 2012, 2013; Gofford et al. 2014, 2015; Reeveset al. 2014).The quasar PDS 456 (at redshift z e ≈ . ) has thebest-studied example of a powerful, relativistic X-ray UFO(e.g., Reeves et al. 2003, 2009, 2014, 2016; Nardini et al.2015; Gofford et al. 2014; Matzeu et al. 2017). It is the mostluminous quasar in the local universe, with bolometric lu-minosity L ∼ ergs s − and estimated black hole mass ∼ ( − ) × M (cid:12) that together indicate an accretion raterelative to Eddington that is L / L Edd (cid:38) . and perhaps nearunity (Nardini et al. 2015). The X-ray absorber is complexand highly variable, with two main components. The maincomponent measured via Fe K-shell absorption has speeds inthe range v ∼ . –0.34 c , very high degrees of ionisation fea-turing Fe xxv and Fe xxvi , and large total column densities log N H ( cm − ) (cid:38) (Reeves et al. 2014; Gofford et al. 2014,2015; Nardini et al. 2015). Its radial distance from the blackhole is estimated at a few hundred gravitational radii (basedon absorber variability, e.g., Nardini et al. 2015). The kineticpower of this outflow is also remarkably large, crudely ∼ ∼ . –0.27 c , lower ionisations, and col-umn densities in the range log N H ( cm − ) ∼ to 23 (Reeveset al. 2016). There is evidence for time-variable covering frac-tions in this absorber that might be indicative of small denseclumps or substructures embedded in the overall X-ray out-flow (Matzeu et al. 2016).PDS 456 also has signatures of outflow in the UV. AUV spectrum obtained in 2000 by O’Brien et al. (2005) us-ing the Space Telescope Imaging Spectrometer (STIS) onboard the Hubble Space Telescope (HST) revealed a highlyblueshifted C iv broad emission line at speeds near v ∼ km s − plus a single BAL that might plausibly be Ly α orN v at v ∼ . c or ∼ c , respectively. These UV featuresprovide further evidence for exotic mass loss from PDS 456across a wide range of spatial scales, from the X-ray UFO that is believed to originate near the black hole at radiiof ∼ – r g (gravitational radii, corresponding to ∼ iv broad emission-line region that weplace crudely at ∼ M (cid:12) black hole). TheUV BAL is an important component to this outflow picture,but its measured speed and physical nature depend criticallyon the line identification.In this paper, we reexamine the UV BAL in PDS 456with the main result that it is likely to be C iv λ ≈ . c . This would be the fastest UV outflow line ever re-ported but similar in speed to the X-ray UFO in this quasar.Throughout this paper, we adopt a redshift for PDS 456of z e = . based on the emission line [Fe ii ] 1.6435 µ m measured by Simpson et al. (1999). We describe twoarchival HST spectra of PDS 456 in Section 2 below. Section3 presents our analysis and measurements of the observedBAL, comparisons to photoionisation models, comparisonsto UV BALs in other quasars, and a discussion of the plau-sible BAL identifications. Section 4 presents a summary anddiscussion of the results. Figure 1 shows the spectrum of PDS 456 obtained byO’Brien et al. (2005) in 2000 using HST-STIS with theG140L and G230L gratings. We obtained this spectrum andanother HST spectrum measured in 2014 (described below)from the Mikulski Archive for Space Telescopes (MAST).We present them here without further processing. The STISG140L grating provided wavelength coverage from 1137 ˚Ato 1715 ˚A at resolutions ranging from ∼
310 km s − to ∼ − , while STIS G230L covered 1580 ˚A to 3148 ˚A atresolutions from ∼
640 km s − to ∼
300 km s − (see O’Brienet al. 2005 for more details). The spectrum plotted in Fig-ure 1 is combined from the two gratings by masking out theextreme ends of the wavelength coverage to avoid excessivenoise and then calculating variance-weighted average fluxesat the remaining wavelengths of overlap.The UV spectrum of PDS 456 is substantially reddenedby dust in our Galaxy (due to the quasar’s sky position nearthe Galactic plane). Previous studies based on visible spectraand photometry indicate that the visual extinction towardsPDS 456 is A V ∼ . magnitudes, corresponding to selectiveextinction E ( B − V ) ∼ . for a standard Galactic reddeningcurve (Torres et al. 1997; Simpson et al. 1999). We obtainan independent estimate of the reddening by fitting the UVcontinuum with a single power law modified by the Galacticextinction curve from Cardelli et al. (1989) with R V = . .The fit is constrained by the median flux in wavelength win-dows between ∼ ∼ E ( B − V ) = . and powerlaw index α λ = − . (for F λ ∝ λ α λ ). An important featureof Galactic extinction at these wavelengths is the “bump”at ∼ E ( B − V ) and the power law slope are well constrained by the We note, however, that the observed spectrum is poorly fit atMNRAS000
300 km s − (see O’Brienet al. 2005 for more details). The spectrum plotted in Fig-ure 1 is combined from the two gratings by masking out theextreme ends of the wavelength coverage to avoid excessivenoise and then calculating variance-weighted average fluxesat the remaining wavelengths of overlap.The UV spectrum of PDS 456 is substantially reddenedby dust in our Galaxy (due to the quasar’s sky position nearthe Galactic plane). Previous studies based on visible spectraand photometry indicate that the visual extinction towardsPDS 456 is A V ∼ . magnitudes, corresponding to selectiveextinction E ( B − V ) ∼ . for a standard Galactic reddeningcurve (Torres et al. 1997; Simpson et al. 1999). We obtainan independent estimate of the reddening by fitting the UVcontinuum with a single power law modified by the Galacticextinction curve from Cardelli et al. (1989) with R V = . .The fit is constrained by the median flux in wavelength win-dows between ∼ ∼ E ( B − V ) = . and powerlaw index α λ = − . (for F λ ∝ λ α λ ). An important featureof Galactic extinction at these wavelengths is the “bump”at ∼ E ( B − V ) and the power law slope are well constrained by the We note, however, that the observed spectrum is poorly fit atMNRAS000 , 1–11 (2015) oes PDS 456 have a UV outflow at 0.3c? F l u x ( − e r g s s − c m − A − ) OIII]HeIICIVSiIVSiIIOI CIINVLy α OVI CIV at 0.3c?
Galactic Abs.
ReddeningCorrected F l u x ( − e r g s s − c m − A − ) Galactic Abs.
CIV at 0.3c?Galactic Ly α Observed
Figure 1.
HST STIS spectrum of PDS 456 from 2000 plot-ted as observed (bottom panel) and reddening-corrected usinga Galactic extinction curve with E ( B − V ) = . (top panel).The grey dotted curves are the corresponding error spectra. Thered dashed curves are an approximate fit to the continuum usinga power law with index α λ = − . shown with (bottom panel)and without (top) Galactic reddening. The red dotted curves inthe bottom panel show the same power law reddened alterna-tively by E ( B − V ) = . and 0.48. The BAL we attribute toC iv at v ≈ . c is labeled above the spectrum at 1346 ˚A. Somebroad emission-line wavelengths are marked across the top. Galac-tic absorption lines (including strong damped Ly α at 1216 ˚A) aremarked by green vertical lines near the bottom. See Section 2. data because Galactic reddening curves with a strong 2175˚A bump suppress the flux at both ends of the spectral cov-erage shown in Figure 1. The top panel in this figure showsthe reddening-corrected spectrum of PDS 456 together withthe best-fit power law.Figure 2 shows a previously-unpublished spectrum ofPDS 456 obtained in 2014 with HST-COS using the G140Lgrating (PI: O’Brien). This spectrum (shown by the redcurve) is plotted on top of the HST-STIS spectrum (blackcurve) from Figure 1 with no reddening corrections. It iscombined from separate exposures totaling 5186 seconds.The spectral resolution ranges from roughly 200 km s − to100 km s − from blue to red across the wavelengths shown wavelengths (cid:38) ii emission lines that are knownto be strong in PDS 456 (Simpson et al. 1999). We do not investi-gate this further because our goal is simply to define a continuumfor analysis of the outflow lines at shorter wavelengths (Section3). F l u x ( − e r g s s − c m − A − ) HeIICIVSiIVSiII OI CIIGeo.OI NVLy α OVI CIV if NV, Ly α CIV at 0.2c?OVI if NV, Ly α CIV at 0.3c?
Figure 2.
HST spectra of PDS 456 obtained with STIS in 2000(black curve) and COS in 2014 (red) plotted at observed wave-lengths, uncorrected for reddening. The red and grey dottedcurves are the corresponding error spectra (1 σ uncertainties perpixel). The STIS spectrum is scaled vertically by a factor ∼ iv λ c is labeledbelow the spectrum at ∼ vi λ iv BALs if the observed BAL is at-tributed, instead, to N v λ c or Ly α at 0.06 c (e.g., “ifN v , Ly α ”). We tentatively identify another weak BAL at ∼ iv at 0.19 c . Dashed blue vertical lines mark the positions ofGalactic absorption features. The narrow emission spike at 1304˚A is geocoronal O i . See Figure 1 and Section 2 for additionalnotes. in Figure 2. A gap in the COS wavelength coverage from ∼ ∼ α ab-sorption/emission. The blue arrows in Figures 1 and 2 markthe positions of observed or expected absorption lines dis-cussed in Section 3 below The HST-STIS spectrum in Figure 1 clearly shows a BALat ∼ τ v = τ o e − v / b (1)where τ o is the line-center optical depth, v is the velocityshift from line center, and b is the doppler parameter thatsets the line width. The BAL fit shown by the magenta curvein Figure 3 yields τ o = . ± . , b = ± km s − , andobserved line-center wavelength λ o = . ± . ˚A (wherethe errors quoted are 1 σ uncertainties returned by the line-fitting routine). The full width at half minimum of this fittedprofile is FWHM ≈ ± km s − . In the quasar frame,the fit has rest equivalent width REW = . ± . ˚A.Our study concerns identification of the ∼ MNRAS , 1–11 (2015)
F. Hamann et al. N o r m a li z e d F l u x CIV at 0.3c?OVI if NV, Ly α CIV if NV, Ly α CIVSiIVSiII OI CIINVLy α OVI Galactic AbsorptionGalactic Ly α Figure 3.
HST-STIS spectrum of PDS 456 in the observed frame,normalized by the continuum fit shown in Figure 1. The smoothmagenta and orange curves show our fit to the observed BAL at1346 ˚A and then that fit transposed to the wavelengths of O vi and C iv assuming the measured BAL is Ly α (magenta) or N v (orange). The weaker profiles drawn at the C iv positions show thetransposed fit at 1/3 its measured strength to illustrate approx-imate upper limits to the actual absorption there. The absenceof absorption at the transposed fit positions argues against theLy α and N v identifications for the observed BAL (Section 3.5and Section 3.6). Other labels are the same as Figures 1 and 2.Narrow Galactic absorption lines of O i λ ii λ Ly α ≈ . c (18,000 km s − ). An-other is N v λ ≈ . c . These identifications seemreasonable because the observed BAL is on the blue side ofthe Ly α –N v emission-line blend and the inferred velocitiesare in a normal range for BAL outflows in other quasars.However, we will argue below (Section 3.5 and Section 3.6)that the Ly α or N v identifications are problematic becausethey are not accompanied by C iv and O vi λ iv andO vi BALs . At the C iv positions, the transposed BAL fitis shown twice – once at full strength and once at 1/3 ofits measured strength at 1346 ˚A. We estimate that the 1/3scalings represent approximate upper limits to C iv BALsthat might accompany a N v or Ly α BAL at 1346 ˚A. Wewill argue below that the absence O vi and C iv absorptionat these wavelengths suggests that the BAL at ∼ iv λ ≈ . c .We test these line identifications using photoionisationmodels (Section 3.3) and comparisons to BALs/mini-BALsobserved in other quasars (Section 3.4). First we estimateionic column densities for each possible identification. Theseestimates follow from the fitted Gaussian optical depth pro-file by N i = m e c √ π e b τ o f λ r (2)where N i is the ionic column density, f and λ r are the oscil-lator strength and rest-frame wavelength of the transition,and we assume implicitly that the ground-state column den-sity equals the ionic column density. Thus we find that, if the observed BAL is Ly α , it corresponds to a neutral hydrogencolumn density of log N HI ( cm − ) ≈ . . If the BAL is N v orC iv , then the column density is log N ( N v )( cm − ) ≈ . or log N ( C iv )( cm − ) ≈ . , respectively. The uncertainties inthese estimates depend mainly on the continuum placementused for the BAL profile fit. Experiments with different con-tinua suggest that the maximum uncertainties are roughly0.1 dex.It is important to note that the line optical depth andcolumn densities derived above are only lower limits becauseBAL outflows often exhibit saturation with partial line-of-sight covering of the background light source. This can leadto weak/shallow absorption-line troughs even if the opti-cal depths are large (e.g., Hamann 1998; Hamann & Sabra2004; Arav et al. 2005; Capellupo et al. 2017; Moravec et al.2017, and refs. therein). In one recent study, Herbst et al.(in prep.) used median composite spectra of BAL quasarsfrom the Baryon Oscillation Sky Survey (BOSS, Dawsonet al. 2013; Ross et al. 2012, part of the Sloan Digital SkySurvey-III, SDSS-III, Eisenstein et al. 2011) to show that thelow-abundance doublet P v λ v , Si iv λ vi λ ∼ iv λ v λ vi λ typ-ically saturated and the observed depths of the BAL troughsare controlled mainly by the line-of-sight covering fractions. In addition to the UV BAL at 1346 ˚A, there are two otherfeatures in the 2000 HST-STIS spectrum that point to ex-treme outflows in PDS 456. First, as noted previously byO’Brien et al. (2005), the broad C iv emission line is ex-tremely blueshifted. Our simple Gaussian fit to this fea-ture shown in Figure 3 indicates that the line centroid isblueshifted by ± km s − (consistent with the O’Brienet al. 2005 measurement). This fit also yields a rest equiva-lent width of REW = . ± . ˚A and a large velocity widthof FWHM = ± km s − (not corrected for the ∼
500 km s − doublet separation in C iv ). The low-ionisationemission lines O i λ ii λ iv , but they are clearly lessblueshifted. We crudely estimate their centroid blueshifts byvisual inspection to be ± km s − .The large C iv emission-line blueshift identifies anoutflow-dominated broad emission-line region with an un-usually high outflow speed. At v ≈ ± km s − , theC iv blueshift is outside of the range of values reported byCoatman et al. (2016) for a sample of ∼ iv blueshift in PDS 456 is in the upper ∼ λ iv blueshift in a futurepaper. Here we note simply that large blueshifts are known MNRAS000
500 km s − doublet separation in C iv ). The low-ionisationemission lines O i λ ii λ iv , but they are clearly lessblueshifted. We crudely estimate their centroid blueshifts byvisual inspection to be ± km s − .The large C iv emission-line blueshift identifies anoutflow-dominated broad emission-line region with an un-usually high outflow speed. At v ≈ ± km s − , theC iv blueshift is outside of the range of values reported byCoatman et al. (2016) for a sample of ∼ iv blueshift in PDS 456 is in the upper ∼ λ iv blueshift in a futurepaper. Here we note simply that large blueshifts are known MNRAS000 , 1–11 (2015) oes PDS 456 have a UV outflow at 0.3c? to correlate with small emission-line REWs and intrinsicallyweak X-ray emissions (compared to other quasars/AGN atsimilar luminosities, e.g., Leighly 2004; Leighly et al. 2007;Wu et al. 2011, 2012; Richards et al. 2011; Luo et al. 2015).In addition, the work by Coatman et al. (2016) supportsspeculation in the ERQ studies that large blueshifts andother prominent outflow features are related to high accre-tion rates (high Eddington ratios) in the quasars.Another tentative outflow feature in the 2000 HST-STISspectrum is a weak BAL marginally detected at ∼ ∼ iv absorption at 1549 ˚A and it appears to bemuch narrower than the BAL candidate at ∼ λ o ≈ ± ˚A, FWHM ≈ ± km s − , and REW ≈ . ± . ˚Ain the quasar frame. If this BAL is real, the only plausibleidentification is C iv at v ∼ . c (The leading alternative,S iv λ ∼ . c is ruled out by the absence of C iv absorption at the same velocity shift. See Rodr´ıguez Hidalgoet al. 2011; Hamann et al. 2013). We use the photoionisation and spectral synthesis codeCloudy version 17.00 (Ferland et al. 2013, 2017) to predictabsorption-line strengths for different physical conditionsthat might produce the observed BAL at ∼ n H = cm − . The specific density has no bear-ing on our results (e.g., Hamann 1997). The metallicity canaffect mainly the metal line strengths relative to the Lymanlines. We choose twice-solar metallicity to be crudely consis-tent with other studies of the outflows and broad emission-line environments of luminous quasars (Hamann & Ferland1999; Dietrich et al. 2002; Warner et al. 2004; Nagao et al.2006; Gabel et al. 2005a, e.g.,). The clouds are irradiated bya standard quasar ionising spectrum consistent with mea-surements of PDS 456 (e.g., Nardini et al. 2015). This spec-trum is defined by power laws across optical-UV and X-raywavelengths with slopes of α uv = − . and α x = − . , respec-tively (for f ν ∝ ν α ). The two power laws are joined smoothlyin the far-UV by an exponential Wien function with temper-ature T = 250,000 K. The relative strengths of the UV andX-ray spectral segments are set by a two-point power-lawindex between 2500 ˚A and 2 keV equal to α ox = − . (Strat-eva et al. 2005; Steffen et al. 2006). This spectrum is definedin Cloudy by the command: AGN T=250000K, a(ox)=-1.8,a(uv)=-0.5, a(x)=-1.3 . The flux of hydrogen ionising ra-diation incident on the model clouds is set by the ionisationparameter, U ≡ Q H π cR n H (3)where R is the distance from the quasar light source and Q H is the total emitted luminosity of hydrogen-ionising photons( Here we construct a sample of outflow quasars from SDSS-III/BOSS that can be useful analogues to test the BALidentification in PDS 456. We specifically use data fromthe BOSS quasar catalog for data release 12 (DR12, Pˆariset al. 2017) to select quasars with the following properties:1) We require that the BOSS spectra have signal-to-noiseratios SNR > z e > . to place the outflow linesof C iv , N v and Ly α within the BOSS wavelength cover-age at speeds up to at least ∼ c . Nearly 2/3 of the se-lected quasars have z e > . such that O vi λ iv outflow lines(BALs or mini-BALs) characterized by balnicity index BI < − and absorption index AI > − , bothmeasured at > σ confidence (see Pˆaris et al. 2017, and refs.therein for definitions of BI and AI). These parameter limitsare designed to exclude strong/deep C iv BALs (with largeBI) as well as very narrow C iv lines with FWHMs < severalhundred km s − (with small AI) that do not resemble the ob-served feature in PDS 456. Note that broad outflow lines arerecorded in the BOSS quasar catalog only at velocity shifts v (cid:46) . c . 4) The minimum velocity recorded on the red sideof the BAL/mini-BAL troughs is v > − (as speci-fied by the AI integration limit vmin_civ_450 > ∼ that can have different pro-files and other properties compared to higher-velocity BALsand mini-BALs. The specific value of 6000 km s − is a com-promise that requires a substantial velocity shift while stillmaintaining a large sample size. Finally, 5) we reject a smallfraction of the quasars ( ∼
10 percent) that have damped Ly α absorption (DLAs) or strong sub-DLAs at the wavelengthsof interest (based on our own visual inspections of the spec-tra).The final sample includes 641 quasars at median red-shift (cid:104) z e (cid:105) = . and median absolute i -band magnitude (cid:104) M i (cid:105) = − . . We visually inspect all of the quasar spectra toassess the strengths of prominent outflow lines for compar-ison to PDS 456. At these redshifts and velocity shifts, theN v , Ly α , and O vi outflow lines are inevitably blended withunrelated absorption features in the Ly α forest. However, inmost cases, the outflow lines are clearly discernible becausethey are much broader than the forest features. Figure 4shows spectra of four quasars we find to have the strongestwell-measured N v outflow lines relative to C iv in our BOSSsample. These extreme cases with large N v /C iv are usefulfor our discussions of the BAL identification in PDS 456(e.g., Section 3.6).It is also useful to consider the median outflow line prop-erties in the BOSS sample. Thus we construct a mediancomposite spectrum in the outflow absorber frame for all641 quasars. This composite “averages out” the Ly α forestcontamination to reveal the typical outflow line strengthsand profiles across the sample. The result is shown inFigure 5. We derive an absorber redshift for each quasarfrom the wavelength of minimum flux between the limitsof the C iv AI integration (e.g., between vmin_civ_450 and vmax_civ_450 ) in smoothed versions of the BOSS spectra.The composite is then constructed by normalizing each spec-trum to a continuum flux near 1700 ˚A and then shiftingto the absorber frame before calculating the median. We
MNRAS , 1–11 (2015)
F. Hamann et al. F l u x CIII OVI SIV Ly α NVSiII OI CII SiIV NIV] SiIICIV
J011210.66-040450.1 z e =2.53 BI=758 AI=2294 F l u x CIII OVI SIV Ly α NVSiII OI CII SiIV NIV] SiIICIV
J015229.73+064309.2 z e =3.67 BI=946 AI=1772 F l u x CIII OVI SIV Ly α NVSiII OI CII SiIV NIV] SiIICIV
J111444.60+642921.3 z e =3.02 BI=494 AI=1396
900 1000 1100 1200 1300 1400 1500Rest Wavelength (A)0.00.51.01.52.02.53.03.5 F l u x CIII OVI SIV Ly α NVSiII OI CII SiIV NIV] SiIICIV
J163909.10+282447.1 z e =3.82 BI=0 AI=1384 Figure 4.
Normalized spectra in the quasar frame of four outflowquasars from our BOSS sample with the strongest well-measuredN v absorption lines relative to C iv . Blue arrows mark the broadoutflow absorption lines of O vi , N v , and C iv from left to right,respectively. N v and O vi suffer from varying amounts of con-tamination in the Ly α forest (e.g., in the bottom panel there isLyman limit absorption at wavelengths (cid:46)
912 ˚A and the deep-est part of the broad N v trough is enhanced and offset from theabsorption minima in C iv and O vi due to Ly α forest contam-ination). Prominent broad emission lines are labeled across thetop. Also listed are the quasar names, emission-line redshifts, z e ,and absorption indices, BI and AI, from the BOSS DR12 quasarcatalog. See Section 3.4 remove broad emission lines from the final outflow quasarcomposite spectrum by dividing by another composite spec-trum of ∼ Mi ) but without broad outflow lines intheir spectra. The redshifts of these non-outflow quasars arerandomly displaced to match the distribution of shifts usedto place the outflow quasars in the absorber frame. A simpledivision then yields the final composite spectrum shown inFigure 5. See Baskin et al. (2013), Baskin et al. (2015), andHerbst et al. (in prep.) for more discussion of this procedure.The main result from Figure 5 is that the O vi λ v λ iv λ α (in the blue wing of N v ) is substantiallyweaker. This represents the typical situation for luminousquasars with BALs/mini-BALs crudely similar to the BALin PDS 456. Another interesting result in Figure 5 is the sig-nificant presence of P v λ N o r m a li z e d F l u x CIIINIIILy β OVI SIVNII PV CIII* Ly α NVSiIIISiII SiII OI CII SiIV SiIICIV1000 1100 1200 1300 1400 1500 1600Rest Wavelength (A)10 N u m b e r Q S O s Figure 5.
Top panel:
Normalized composite spectrum in theabsorber frame for 641 BOSS quasars selected to have C iv BALs/mini-BALs with moderate strengths and minimum troughvelocities > − (Section 3.4). The dashed blue verticallines with labels across the top mark absorption lines that are ormight be present. The dashed red horizontal line marks the unab-sorbed continuum. Bottom panel:
Number of quasars contributingto the composite at each wavelength. doublets with ∼ vi , P v , and Si iv . This sup-ports the claim by Herbst et al. (in prep.) that BAL/mini-BAL outflows often have large total column densities andsaturated absorption in all of the prominent lines. α Identification
Figure 6 shows theoretical line-center optical depths fromour Cloudy simulations (Section 3.2) for lines that shouldaccompany a Ly α BAL at 1346 ˚A in PDS 456. In particu-lar, the solid curves show the optical depths for lines with b = km s − formed in clouds with a range of ionisationparameters, U , but with neutral hydrogen column densitiesheld fixed at the derived value of log N HI ( cm − ) = . (Sec-tion 3.1). These predictions should be considered lower limitsbecause N HI derived from the data is a lower limit. Holding N HI constant in the calculations leads to model clouds thathave larger total hydrogen column densities, N H , at larger U . The log N H values for each log U are shown across the topof the figure.The shaded regions around the curves for C iv , N v , andO vi in Figure 6 illustrate their dependence on the specificvalues of N HI and the turbulent velocity, v turb , used in thecalculations. The solid curves for these lines (and the solidand dashed curves for C iii λ iii λ iv λ log N HI ( cm − ) = . andv turb = km s − . In the top panel, the shaded regionsshow the range of optical depths corresponding to the maxi-mum uncertainty of ± . dex in our log N HI measurement(see Section 3.1, where the smaller/larger N HI value setsthe lower/upper envelope to the shaded curves). The opticaldepth changes in response to these changes in N HI are non-linear at large U (and large N H ) because of radiative shield-ing in the far-UV that affects each ion differently. Theseshielding effects maintain significant optical depths in themoderate ions C iv , N v , and O vi at a fixed N HI and large U . In other calculations (not shown), we determine that thethreshold for important shielding effects at large U is roughly MNRAS000
Figure 6 shows theoretical line-center optical depths fromour Cloudy simulations (Section 3.2) for lines that shouldaccompany a Ly α BAL at 1346 ˚A in PDS 456. In particu-lar, the solid curves show the optical depths for lines with b = km s − formed in clouds with a range of ionisationparameters, U , but with neutral hydrogen column densitiesheld fixed at the derived value of log N HI ( cm − ) = . (Sec-tion 3.1). These predictions should be considered lower limitsbecause N HI derived from the data is a lower limit. Holding N HI constant in the calculations leads to model clouds thathave larger total hydrogen column densities, N H , at larger U . The log N H values for each log U are shown across the topof the figure.The shaded regions around the curves for C iv , N v , andO vi in Figure 6 illustrate their dependence on the specificvalues of N HI and the turbulent velocity, v turb , used in thecalculations. The solid curves for these lines (and the solidand dashed curves for C iii λ iii λ iv λ log N HI ( cm − ) = . andv turb = km s − . In the top panel, the shaded regionsshow the range of optical depths corresponding to the maxi-mum uncertainty of ± . dex in our log N HI measurement(see Section 3.1, where the smaller/larger N HI value setsthe lower/upper envelope to the shaded curves). The opticaldepth changes in response to these changes in N HI are non-linear at large U (and large N H ) because of radiative shield-ing in the far-UV that affects each ion differently. Theseshielding effects maintain significant optical depths in themoderate ions C iv , N v , and O vi at a fixed N HI and large U . In other calculations (not shown), we determine that thethreshold for important shielding effects at large U is roughly MNRAS000 , 1–11 (2015) oes PDS 456 have a UV outflow at 0.3c? L o g L i n e O p t i c a l D e p t h s
18 19 20 21 22 23
Log N H (cm − ) if Ly α : v turb = 1000 km/s log N HI (cm − ) = 15.35 ± Log U L o g L i n e O p t i c a l D e p t h s if Ly α : log N HI (cm − ) = 15.35 v turb = 100, 1000, 5135 km/s OVINVCIVCIIISiIII SiIV Figure 6.
Line-center optical depths versus ionisation param-eter, log U (bottom axis), and total hydrogen column density, log N H (top axis) predicted by Cloudy simulations to accompanya Ly α BAL like the observed feature in PDS 456 (with τ o = . and b = km s − , Section 3.1). The dashed and dark solidcurves are the same in both panels, representing model cloudswith log N HI ( cm − ) = . and v turb = km s − . Shaded re-gions around the C iv , N v , and O vi curves show the ranges ofoptical depths that result from varying log N HI by ± turb from 100 km s − to its max-imum value, 5135 km s − , set by the observed BAL width (bot-tom panel). These predictions show that the Ly α BAL identifica-tion would require log U < − and uniquely low column densities, log N H ( cm − ) < . , for a BAL outflow (see Section 3.5). log N HI ( cm − ) (cid:38) . . This is 0.55 dex lower than our lowerlimit to the H i column density inferred from the observedBAL (Section 3.1), and therefore the predictions in Figure6 that result from shielding at large U should be applicable.Another parameter that can affect radiative shielding isthe velocity dispersion inside the cloud, which we character-ize by the turbulence velocity v turb . Large internal velocitydispersions can enhance the shielding by blending togethernumerous absorption lines in the far-UV, which in turn af-fects the ionisation structure deep in the cloud. The shadedregions around the C iv , N v , and O vi curves in the bottompanel of Figure 6 depict a range of results corresponding tov turb = km s − , which yields negligible line shielding (thebottom envelope of the shaded regions), up to v turb = km s − , which yields the maximum line shielding allowed bythe observed BAL width (the upper envelope of the shaded regions). We adopt an intermediate value v turb = km s − as our fiducial case (here and in all of our calculations be-low) because it yields conservatively small amounts of lineshielding and, therefore, conservatively small predicted lineoptical depths in C iv , N v , and O vi at large U .The main result from Figure 6 is that a Ly α BAL shouldbe accompanied by significant broad C iv and O vi absorp-tion lines over a wide range of normal BAL outflow physicalconditions. It is important to keep in mind that the predictedstrengths of the C iv and O vi lines are lower limits basedon a lower limit on N HI that follows from the assumption ofno partial covering. Smaller values of N HI are not allowedbecause they cannot produce the observed BAL attributedto Ly α . Larger N H requiring larger N H (at a given U ) wouldproduce stronger C iv and O vi absorption. The absence ofthese lines in the PDS 456 spectrum (Figures 2 and 3) meansthat the only conditions consistent with the Ly α BAL iden-tification are low degrees of ionisation, log U < − , and verylow total column densities, log N H ( cm − ) < . that are (cid:38) typical BAL outflowshave log N H ( cm − ) (cid:38) . (see also Sections 1 and 3.4).Another problem for the Ly α identification is that aLy α -only BAL system would be, to our knowledge, unprece-dented among observed BAL outflows. Figure 3 shows thatthe C iv and O vi BALs expected to accompany Ly α inPDS 456 would need to be <
33 percent the strength of Ly α to avoid detection. This is very different from the typical sit-uation illustrated by Figure 5 and by other BAL compositespectra (Baskin et al. 2013, Herbst et al., in prep.) wherethe C iv and O vi lines that are ∼ ∼ stronger than Ly α . BAL surveys based on C iv BAL detections (e.g.,Trump et al. 2006; Gibson et al. 2009; Pˆaris et al. 2017) arenaturally biased toward stronger C iv lines and they cannotdetect Ly α -only BAL systems if they exist. However, ourvisual inspections of the 641 outflow quasar spectra in ourBOSS sample (Section 3.4) do not find any instances of C iv BALs weaker than, or even comparable to, the correspond-ing Ly α absorption.A third argument against the Ly α BAL identificationstems from the weakening of the observed BAL between theHST observations in 2000 and 2014 (Figure 2). Variabilityin BALs and other outflow lines can be attributed to out-flow clouds moving across our lines of sight or to changesin the ionisation caused by changes in the incident ionisingflux (cf., Filiz Ak et al. 2013; Capellupo et al. 2013, 2014;Misawa et al. 2014a; Grier et al. 2015; Arav et al. 2012, 2015;Rogerson et al. 2016; Moravec et al. 2017, and refs. therein).The evidence for BAL saturation discussed in Section 3.1and Section 3.4 might favor crossing clouds in many situa-tions. However, if the BAL changes in PDS 456 were causedby ionisation changes and the ionisation was initially lowto produce Ly α absorption without accompanying C iv andO vi lines, then the only way to make Ly α weaker at a fixed N H is with larger ionisation parameters that should producestronger C iv and O vi absorption. The continued absence ofthese accompanying lines in the 2014 (Figure 2) thus makesthe Ly α identification more difficult because the values of N H and U in 2000 would need to be several times lower thanthe already-low upper limits deduced from Figure 6. MNRAS , 1–11 (2015)
F. Hamann et al. L o g L i n e O p t i c a l D e p t h s OVI CIVCIII Ly α SiIV τ (NV) = 0.35 log N(NV) (cm − ) = 15.62 L o g N H ( c m − ) Log U L o g L i n e O p t i c a l D e p t h s NeVIII MgXOVI NVCIIILy α Ly α SiIV τ (CIV) = 0.35 log N(CIV) (cm − ) = 15.44 L o g N H ( c m − ) Figure 7.
Line-center optical depths (black curves) and totalhydrogen column densities, log N H (red curve), versus ionisationparameter, log U , predicted by Cloudy to accompany the observedBAL (with τ o = . and b = km s − ; Section 3.1) if it isattributed to N v (top panel) or C iv (bottom panel). The ioniccolumn densities held fixed in the model clouds are shown in theupper left of each panel. These predictions and existing spectra(Figures 1–3) rule out low degrees of ionisation for the N v BALsystem, and they can guide future observations to test both BALidentifications further (see Section 3.6 and Section 3.7). v λ The top panel in Figure 7 shows Cloudy predictions for theline-center optical depths that should accompany the ob-served BAL if it is N v λ turb = km s − (to include moderate amounts of line shielding) but line op-tical depths derived for b = km s − . The total columndensities in the model clouds are adjusted to yield an N v column density, log N ( N v )( cm − ) = . , fixed to the valueinferred from the observed BAL (Section 3.1). The resultingoptical depths are again only lower limits because N ( N v ) isa lower limit.The observed BAL at 1346 ˚A is compatible with theN v identification only if the accompanying C iv line is un-detectable. This requires a ratio of N v / C iv absorption linestrengths (cid:38) (Section 3.1, Figure 3). Figure 7 shows thatmoderate-to-low outflow ionisations with log U (cid:46) , corre-sponding to total columns log N H ( cm − ) (cid:46) . , are firmlyruled out for the N v BAL identification by the predictionsfor strong C iv absorption. At higher ionisations, the pre-dicted optical depth ratio hovers around τ ( N v )/ τ ( C iv ) ∼ –5 (Figure 7), which is marginally consistent with the ob-served spectrum. However, any significant partial covering effects would push the observed N v /C iv line depth ratiotoward unity, such that C iv should be detectable and theN v BAL identification is ruled out.Figures 4 and 5 show more directly that absorption lineratios N v /C iv (cid:38) needed for the N v BAL identificationare, at best, extremely rare in observed quasar outflows. Inparticular, the composite spectrum (Figure 5) shows thattypical weak-to-moderate BAL/mini-BAL systems have amedian line depth ratio of N v /C iv ∼ ± . . Individual out-flows can have larger N v /C iv ratios. Measurements in theliterature of BALs/mini-BALs at speeds up to nearly 0.2 c reveal N v /C iv depth ratios in the range ∼ ∼ v /C iv (cid:38) . Figure 4 shows the mostextreme well-measured examples. It is difficult to assess theN v /C iv line ratios quantitatively in these spectra due toblending problems in the Ly α forest. However, the most ex-treme cases shown in Figure 4 also appear to be inconsistentwith the N v BAL identification in PDS 456.Another constraint on the N v BAL identification is thatit should be accompanied by strong O vi absorption at ob-served wavelength ∼ vi line wing that could be measurable if this linewas present (Figure 3). The 2014 COS spectrum does coverthese O vi wavelengths and clearly shows no signs of absorp-tion there (Figure 2), but the BAL at 1346 ˚A was also muchweaker in this spectrum. Thus the constraints provided byabsence of O vi in existing spectra are weak, but they alsodo not support the N v BAL identification. iv λ One argument favoring the C iv BAL identification inPDS 456 is that it is readily compatible with existingspectra (Figures 1–3). It has none of the problems de-scribed above for Ly α and N v because no other lines areexpected within the wavelength coverage. Observations ofhigh-velocity BALs/mini-BALs in other quasars indicatethat the outflow ionisations are generally high and that themost prominent lines accompanying C iv should be N v andO vi (Section 3.4, Section 3.6, Hamann et al. 2013). Thebottom panel in Figure 7 shows specific theoretical predic-tions for the optical depths in these and other lines thatshould accompany the observed BAL if it is C iv . The mostobservationally-accessible lines in this plot are Ly α at lowionisations and N v at high ionisations, both at predictedobserver-frame wavelengths near ∼ We need to acknowledge here that hypothetical quasars withstrong N v BALs but negligible C iv absorption cannot appear inour BOSS sample because it relies on C iv line detections (via AIand BI). MNRAS000
Line-center optical depths (black curves) and totalhydrogen column densities, log N H (red curve), versus ionisationparameter, log U , predicted by Cloudy to accompany the observedBAL (with τ o = . and b = km s − ; Section 3.1) if it isattributed to N v (top panel) or C iv (bottom panel). The ioniccolumn densities held fixed in the model clouds are shown in theupper left of each panel. These predictions and existing spectra(Figures 1–3) rule out low degrees of ionisation for the N v BALsystem, and they can guide future observations to test both BALidentifications further (see Section 3.6 and Section 3.7). v λ The top panel in Figure 7 shows Cloudy predictions for theline-center optical depths that should accompany the ob-served BAL if it is N v λ turb = km s − (to include moderate amounts of line shielding) but line op-tical depths derived for b = km s − . The total columndensities in the model clouds are adjusted to yield an N v column density, log N ( N v )( cm − ) = . , fixed to the valueinferred from the observed BAL (Section 3.1). The resultingoptical depths are again only lower limits because N ( N v ) isa lower limit.The observed BAL at 1346 ˚A is compatible with theN v identification only if the accompanying C iv line is un-detectable. This requires a ratio of N v / C iv absorption linestrengths (cid:38) (Section 3.1, Figure 3). Figure 7 shows thatmoderate-to-low outflow ionisations with log U (cid:46) , corre-sponding to total columns log N H ( cm − ) (cid:46) . , are firmlyruled out for the N v BAL identification by the predictionsfor strong C iv absorption. At higher ionisations, the pre-dicted optical depth ratio hovers around τ ( N v )/ τ ( C iv ) ∼ –5 (Figure 7), which is marginally consistent with the ob-served spectrum. However, any significant partial covering effects would push the observed N v /C iv line depth ratiotoward unity, such that C iv should be detectable and theN v BAL identification is ruled out.Figures 4 and 5 show more directly that absorption lineratios N v /C iv (cid:38) needed for the N v BAL identificationare, at best, extremely rare in observed quasar outflows. Inparticular, the composite spectrum (Figure 5) shows thattypical weak-to-moderate BAL/mini-BAL systems have amedian line depth ratio of N v /C iv ∼ ± . . Individual out-flows can have larger N v /C iv ratios. Measurements in theliterature of BALs/mini-BALs at speeds up to nearly 0.2 c reveal N v /C iv depth ratios in the range ∼ ∼ v /C iv (cid:38) . Figure 4 shows the mostextreme well-measured examples. It is difficult to assess theN v /C iv line ratios quantitatively in these spectra due toblending problems in the Ly α forest. However, the most ex-treme cases shown in Figure 4 also appear to be inconsistentwith the N v BAL identification in PDS 456.Another constraint on the N v BAL identification is thatit should be accompanied by strong O vi absorption at ob-served wavelength ∼ vi line wing that could be measurable if this linewas present (Figure 3). The 2014 COS spectrum does coverthese O vi wavelengths and clearly shows no signs of absorp-tion there (Figure 2), but the BAL at 1346 ˚A was also muchweaker in this spectrum. Thus the constraints provided byabsence of O vi in existing spectra are weak, but they alsodo not support the N v BAL identification. iv λ One argument favoring the C iv BAL identification inPDS 456 is that it is readily compatible with existingspectra (Figures 1–3). It has none of the problems de-scribed above for Ly α and N v because no other lines areexpected within the wavelength coverage. Observations ofhigh-velocity BALs/mini-BALs in other quasars indicatethat the outflow ionisations are generally high and that themost prominent lines accompanying C iv should be N v andO vi (Section 3.4, Section 3.6, Hamann et al. 2013). Thebottom panel in Figure 7 shows specific theoretical predic-tions for the optical depths in these and other lines thatshould accompany the observed BAL if it is C iv . The mostobservationally-accessible lines in this plot are Ly α at lowionisations and N v at high ionisations, both at predictedobserver-frame wavelengths near ∼ We need to acknowledge here that hypothetical quasars withstrong N v BALs but negligible C iv absorption cannot appear inour BOSS sample because it relies on C iv line detections (via AIand BI). MNRAS000 , 1–11 (2015) oes PDS 456 have a UV outflow at 0.3c? due to the short wavelength and severe reddening (Section2). The situation is worse for other corroborating lines. Inparticular, the corresponding Si iv λ ∼ α line, and the O vi and Ne viii λ
774 lines, whichshould be strong at high ionisations are at inaccessible ob-served wavelengths ( ∼
899 ˚A and ∼
673 ˚A, respectively) dueto Galactic Lyman limit absorption.A second circumstantial argument favoring the C iv identification is that its velocity v ≈ . c is similar to thespeeds measured for X-ray outflow of PDS 456. The X-rayoutflow is highly variable with multiple velocity components,but the main component identified by Fe K-shell absorptionhas measured velocities in the range v ∼ . –0.34 c with atypical value near ∼ c (Reeves et al. 2014, 2016; Nardiniet al. 2015; Matzeu et al. 2017). The C iv BAL might pro-vide evidence for a physical relationship between the UVand X-ray outflows in PDS 456 (see Section 4 below).A third argument is that relativistic C iv BALs/mini-BALs at speeds approaching v ∼ . c have already beenmeasured in a growing number of luminous quasars (Jan-nuzi et al. 1996; Hamann et al. 1997b, 2013; Rodr´ıguez Hi-dalgo 2008; Rodr´ıguez Hidalgo et al. 2011; Rogerson et al.2016, Rodriguez Hidalgo et al., in prep.). The C iv BAL atv ≈ . c in PDS 456 would set a new speed record for UVoutflows, but it is not so dramatic to be a paradigm shift forour understanding of these outflows. The width and depthof the BAL in PDS 456 is roughly similar to these otherhigh-velocity outflow features. We also note that our ten-tative detection of another weak C iv BAL in PDS 456 atv ∼ . c (Figure 2, Section 3.2) provides additional, albeittentative, evidence that a relativistic UV outflow is presentin PDS 456. The UV spectrum of PDS 456 obtained with HST-STIS in2000 has a distinct BAL at ∼ iv λ ≈ . c , FWHM ≈ km s − ,and minimum optical depth τ o = . (Section 3.1). TheC iv identification rests on its compatibility with existingspectra and the absence of lines that should accompany thealternative identifications, Ly α λ v λ α is compatible with the observed BAL only if the outflowhas a low degree of ionisation, log U < − , and a very lowtotal column density, log N H ( cm − ) < . , that would beunprecedented in BAL outflow studies, e.g., several ordersof magnitude below recent estimates (Section 3.5). The N v identification might be consistent with absence of accom-panying C iv absorption if the gas is highly ionised and itpushes the boundary of observed N v /C iv line strengths be-yond what we find in our comparison of 641 outflow quasarsin BOSS (Section 3.6). However, this situation should pro-duce strong O vi absorption, which is not well constrainedin existing spectra of PDS 456 but it appears to be absent(Figure 3). Thus the N v identification is also strongly dis-favored.The C iv BAL identification has none of these problems(Section 3.7). It would mark the fastest UV outflow line everreported, but its velocity is consistent with the X-ray outflowin PDS 456 (see Section 1 and below) and not dramatically different from the relativistic C iv BALs/mini-BALs alreadymeasured at speeds approaching ∼ c in other quasars (Jan-nuzi et al. 1996; Hamann et al. 1997b, 2013; Rodr´ıguez Hi-dalgo 2008; Rodr´ıguez Hidalgo et al. 2011; Rogerson et al.2016, Rodriguez Hidalgo et al., in prep.). The C iv identifica-tion is also weakly supported by our tentative identificationof an additional C iv BAL feature at v ∼ . c (, Figure 2,Section 3.2). Broad UV outflow lines at speeds near . c aresurely very rare based on the rarity of such lines at ∼ c (see refs. above). However, the incidence of C iv BALs/mini-BALs at v > . c is not known because they have not beensearched for in large quasar surveys like SDSS/BOSS and,in any case, they could easily be missed due to blends withunrelated lines in the Ly α forest (e.g., see the search for P v λ iv BAL outflow in PDS 456 is acritical unknown. The range of ionizations consistent withC iv absorption might favor lower ionizations and larger dis-tances from the black hole than the X-ray outflow, perhapsat radii of order ∼ iv BALs/mini-BALs in other quasars (Hamann et al.2013; Capellupo et al. 2014; Moravec et al. 2017, McGrawet al., submitted). However, larger distances do not neces-sarily produce lower ionisations, e.g., if acceleration causesthe outflow densities to drop faster than the / r behaviourexpected from free expansion at a constant speed. Moreover,the relativistic speed of the C iv BAL at v ∼ . c indicatesthat the UV outflow originated with the X-ray UFO veryclose to the black hole. If the measured flow speeds are sim-ilar to the gravitational escape speed at the launch radius,then the launch point is roughly at r ∼ – r g (Nardiniet al. 2015; Matzeu et al. 2017). Also note that a highly-ionised X-ray outflow launched from this radius and expand-ing freely outward into a fixed solid angle will not necessarilybecome less ionised at larger distances (for C iv absorption)because the / r dilution of the radiation field is balanced bya / r decline in the densities to yield a constant ionisationparameter. Lower ionisations will occur if the inner regionsof the outflow radiatively shield the material downstreamand/or if there are clumps with enhanced densities relativeto the ambient flow. Clumping and shielding can occur tothe same effect at almost any radius, and there is alreadyevidence for dense clumps in the X-ray outflow of PDS 456(based on lower ionizations and partial covering in the softX-ray absorber, Reeves et al. 2016; Matzeu et al. 2016; seealso Gofford et al. 2014; Nardini et al. 2015; Hagino et al.2015).It is therefore an intriguing possibility that the C iv BAL forms directly within, or in close proximity to, the rel-ativistic X-ray outflow. Our Cloudy simulations (Figure 7)show that the C iv BAL in PDS 456 could form over a widerange of physical conditions, including very high ionisationswhere C iv is just a trace constituent. This situation is il-lustrated in Figure 8, which plots the ionisation structurein a single model cloud with log U = . . This cloud reachesthe observed minimum BAL optical depth τ o ( C iv ) = . ,along with τ o ( N v ) = . and τ o ( O vi ) = , at total col-umn density log N H ( cm − ) = . (represented by the un-shaded left-hand portion of Figure 8). This front portion ofthe cloud matches our calculations in the bottom panel ofFigure 7. In this environment, the C iv and N v ion fractionsare everywhere (cid:46) × − and (cid:46) × − , respectively, and the MNRAS , 1–11 (2015) F. Hamann et al.
Log N H (cm − ) L o g I o n F r a c t i o n NeVIIIMgX OVIOVIIOVIII NV CIV C I V H e II Figure 8.
Ionisation fractions versus total hydrogen column den-sity, log N H , in a single model cloud with ionisation parameter, log U = . . This cloud produces the minimum observed BAL opti-cal depth in C iv τ o ( C iv ) = . , at log N H ( cm − ) ≈ . , such thatthe unshaded left-hand region represents the model cloud with log U = . in Figure 7. This environment (at log N H ( cm − ) ≤ . )is highly ionised and optically thick at the bound-free edges ofO vii , O viii , and Ne viii , with only trace amounts of the lowerions C iv , N v , and O vi . The lower ions dominate at larger col-umn densities (grey shaded region) where there is more radiativeshielding, but these lower ionisations are not needed for the C iv BAL in PDS 456. dominant form of oxygen is O vii . This highly-ionsed envi-ronment capable of C iv λ viii λ x λ
615 (see also log U = . Figure 7) and at thebound-free edges of O vii and O viii in soft X-rays.Detailed comparisons between the UV and X-ray out-flows in PDS 456 are beyond the scope of the present study.They are subject to uncertainties caused by the outflow vari-abilities and by unknowns in the shape of the ionising spec-trum and the spatial locations of different outflow compo-nents relative to the ionizing far-UV and X-ray emissionsources. Here we note simply that recent estimates of the X-ray outflow ionisation parameters are roughly in the range log U ∼ . to 4.8 for the K-shell outflow (Gofford et al.2014; Nardini et al. 2015) and log U ∼ . for the soft X-ray absorber (Reeves et al. 2016). The ionisation parametersneeded for the C iv BAL, with upper limit log U (cid:46) . (Fig-ure 7), might occur in dense clumps embedded in the X-rayoutflow if the density enhancements are (cid:38) (cid:38) z ∼ . ), whichhas measured UV BALs ranging in ionization from C iv upto Mg x and Si xii λ
510 (Telfer et al. 1998). The lines in thisquasar appear saturated with shallow troughs that revealion-dependent line-of-sight covering fractions from ∼
15 per-cent in C iv to ∼
50 percent in the higher ions. Telfer et al. The ionisation parameter quoted in the X-ray studies is ξ = L ion / n H r from Tarter et al. (1969), where L ion is the quasar lu-minosity from 1 to 1000 Rydberg. For ξ in units of ergs cm s − and the continuum shape used in our Cloudy simulations (Section3.3), the conversion is log U ≈ log ξ − . . (1998) infer from this a 2-zone outflow that could span adecade or more in ionisation parameter, with the C iv linesforming in small clumps embedded in a more highly-ionisedoutflow medium. Telfer et al. (1998) also note that this out-flow should produce substantial bound-free absorption byO vii and O viii in soft X-rays.The general picture of clumpy outflows has becomecommonplace in quasar outflow studies. High-quality ob-servations of UV outflow lines often provide evidence forclumpy multi-phase outflow structures with a range of cov-ering fractions (Moravec et al. 2017; Misawa et al. 2014b;Hamann et al. 2011; Hamann & Sabra 2004; Gabel et al.2005a; Leighly et al. 2015, 2011, 2009; Misawa et al. 2007;Arav et al. 2008, 2005; de Kool et al. 2002; Ganguly et al.1999). Detailed studies of some bright Seyfert 1 galaxies(e.g., Kaspi et al. 2002; Netzer et al. 2003; Gabel et al. 2005b)clearly demonstrate that the UV and X-ray absorption fea-tures can form together in complex outflows, with indica-tions that the lower-ionisation UV lines identify clumps orfilaments embedded in the X-ray outflow. Clumpy outflowstructures are also predicted by recent numerical simula-tions (Sim et al. 2010; Takeuchi et al. 2013; Waters et al.2017) and by considerations of the radiative forces that cancompress the outflows into small substructures (Stern et al.2014). There are also theoretical arguments requiring denseclumps to moderate the outflow ionizations in the absence ofsignificant radiative shielding (De Kool 1997; Hamann et al.2013).If the C iv -absorbing gas in PDS 456 is indeed embed-ded in the X-ray outflow, it could have important implica-tions for the outflow energetics. It is a well-known problemfor X-ray UFOs (e.g., Tombesi et al. 2011, 2013; Goffordet al. 2015) that their high ionizations lead to low opacitiesand inefficient radiative acceleration (Gofford et al. 2013,2014; Higginbottom et al. 2014). Harnessing the full ra-diative power of the quasar to drive these outflows mightrequire opacities beyond electron scattering at UV/far-UVwavelengths, near the peak of the quasar spectral energydistributions. Matzeu et al. (2017) showed recently that theoutflow speeds in PDS 456 correlate with the variable X-rayluminosity, consistent with radiative acceleration (see alsoSaez & Chartas 2011). If radiative forces are important, thendense clumps with lower ionisations embedded in the X-rayoutflows might be important to boost the opacities for ra-diative driving (Laor & Davis 2014; Hagino et al. 2015). TheC iv BAL at v ≈ . c in PDS 456 could be the first directobservational evidence for this idea.PDS 456 is a remarkable object with powerful accretion-disk outflows launched from a range of at least two decadesin disk radii, from ∼ ∼ iv BAL identification.A search for corroborating N v absorption at ∼ v and Ly α , might be ruled out or confirmed more easilyby searching for O vi absorption near 1122 ˚A (Section 3.6,although the BAL at 1346 ˚A needs to be present for a mean-ingful test). Our team has ongoing programs to extend the MNRAS000
50 percent in the higher ions. Telfer et al. The ionisation parameter quoted in the X-ray studies is ξ = L ion / n H r from Tarter et al. (1969), where L ion is the quasar lu-minosity from 1 to 1000 Rydberg. For ξ in units of ergs cm s − and the continuum shape used in our Cloudy simulations (Section3.3), the conversion is log U ≈ log ξ − . . (1998) infer from this a 2-zone outflow that could span adecade or more in ionisation parameter, with the C iv linesforming in small clumps embedded in a more highly-ionisedoutflow medium. Telfer et al. (1998) also note that this out-flow should produce substantial bound-free absorption byO vii and O viii in soft X-rays.The general picture of clumpy outflows has becomecommonplace in quasar outflow studies. High-quality ob-servations of UV outflow lines often provide evidence forclumpy multi-phase outflow structures with a range of cov-ering fractions (Moravec et al. 2017; Misawa et al. 2014b;Hamann et al. 2011; Hamann & Sabra 2004; Gabel et al.2005a; Leighly et al. 2015, 2011, 2009; Misawa et al. 2007;Arav et al. 2008, 2005; de Kool et al. 2002; Ganguly et al.1999). Detailed studies of some bright Seyfert 1 galaxies(e.g., Kaspi et al. 2002; Netzer et al. 2003; Gabel et al. 2005b)clearly demonstrate that the UV and X-ray absorption fea-tures can form together in complex outflows, with indica-tions that the lower-ionisation UV lines identify clumps orfilaments embedded in the X-ray outflow. Clumpy outflowstructures are also predicted by recent numerical simula-tions (Sim et al. 2010; Takeuchi et al. 2013; Waters et al.2017) and by considerations of the radiative forces that cancompress the outflows into small substructures (Stern et al.2014). There are also theoretical arguments requiring denseclumps to moderate the outflow ionizations in the absence ofsignificant radiative shielding (De Kool 1997; Hamann et al.2013).If the C iv -absorbing gas in PDS 456 is indeed embed-ded in the X-ray outflow, it could have important implica-tions for the outflow energetics. It is a well-known problemfor X-ray UFOs (e.g., Tombesi et al. 2011, 2013; Goffordet al. 2015) that their high ionizations lead to low opacitiesand inefficient radiative acceleration (Gofford et al. 2013,2014; Higginbottom et al. 2014). Harnessing the full ra-diative power of the quasar to drive these outflows mightrequire opacities beyond electron scattering at UV/far-UVwavelengths, near the peak of the quasar spectral energydistributions. Matzeu et al. (2017) showed recently that theoutflow speeds in PDS 456 correlate with the variable X-rayluminosity, consistent with radiative acceleration (see alsoSaez & Chartas 2011). If radiative forces are important, thendense clumps with lower ionisations embedded in the X-rayoutflows might be important to boost the opacities for ra-diative driving (Laor & Davis 2014; Hagino et al. 2015). TheC iv BAL at v ≈ . c in PDS 456 could be the first directobservational evidence for this idea.PDS 456 is a remarkable object with powerful accretion-disk outflows launched from a range of at least two decadesin disk radii, from ∼ ∼ iv BAL identification.A search for corroborating N v absorption at ∼ v and Ly α , might be ruled out or confirmed more easilyby searching for O vi absorption near 1122 ˚A (Section 3.6,although the BAL at 1346 ˚A needs to be present for a mean-ingful test). Our team has ongoing programs to extend the MNRAS000 , 1–11 (2015) oes PDS 456 have a UV outflow at 0.3c? UV wavelength coverage and monitor PDS 456 in the UVand X-rays that will be described in future papers.
ACKNOWLEDGEMENTS
We are grateful to Gary Ferland and the Cloudy devel-opment team for their continued support and public dis-persement of the spectral synthesis code Cloudy. FH alsothanks Nahum Arav, Chris Done, Gerald Kriss, and PaolaRodriguez Hidalgo for helpful discussions. We are gratefulto the referee, Paul Hewett, for helpful comments that im-proved this manuscript. JNR acknowledges the support viaHST grant HST-GO-14477.001-A. EN acknowledges fund-ing from the European Union’s Horizon 2020 research andinnovation programme under the Marie Sklodowska-Curiegrant agreement no. 664931.
REFERENCES
Arav N., et al., 2001, ApJ, 561, 118Arav N., Kaastra J., Kriss G. A., Korista K. T., Gabel J., ProgaD., 2005, ApJ, 620, 665Arav N., Moe M., Costantini E., Korista K. T., Benn C., EllisonS., 2008, ApJ, 681, 954Arav N., et al., 2012, A&A, 544, A33Arav N., et al., 2015, A&A, 577, A37Barlow T. A., Sargent W. L. W., 1997, AJ, 113, 136Baskin A., Laor A., Hamann F., 2013, MNRAS, 432, 1525Baskin A., Laor A., Hamann F., 2015, MNRAS, 449, 1593Borguet B. C. J., Edmonds D., Arav N., Benn C., ChamberlainC., 2012, ApJ, 758, 69Capellupo D. M., Hamann F., Shields J. C., Halpern J. P., BarlowT. A., 2013, MNRAS, 429, 1872Capellupo D. M., Hamann F., Barlow T. A., 2014, MNRAS, 444,1893Capellupo D. M., et al., 2017, MNRAS, 469, 323Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Chartas G., Brandt W. N., Gallagher S. C., Garmire G. P., 2002,ApJ, 579, 169Chartas G., et al., 2009, New Astronomy Review, 53, 128Coatman L., Hewett P. C., Banerji M., Richards G. T., 2016,MNRAS, 461, 647Crenshaw D. M., Kraemer S. B., George I. M., 2003, ARA&A,41, 117Dawson K. S., et al., 2013, AJ, 145, 10De Kool M., 1997, in Arav N., Shlosman I., Weymann R. J.,eds, Astronomical Society of the Pacific Conference SeriesVol. 128, Mass Ejection from Active Galactic Nuclei. p. 233( arXiv:astro-ph/9704230 )Debuhr J., Quataert E., Ma C.-P., 2012, MNRAS, 420, 2221Di Matteo T., Springel V., Hernquist L., 2005, Nature, 433, 604Dietrich M., Hamann F., Shields J. C., Constantin A., Vester-gaard M., Chaffee F., Foltz C. B., Junkkarinen V. T., 2002,ApJ, 581, 912Eisenstein D. J., et al., 2011, AJ, 142, 72Ferland G. J., et al., 2013, Rev. Mex. Astron. Astrofis., 49, 137Ferland G. J., et al., 2017, preprint, ( arXiv:1705.10877 )Filiz Ak N., et al., 2013, ApJ, 777, 168Gabel J. R., et al., 2005a, ApJ, 623, 85Gabel J. R., et al., 2005b, ApJ, 631, 741Ganguly R., Eracleous M., Charlton J. C., Churchill C. W., 1999,AJ, 117, 2594Gibson R. R., et al., 2009, ApJ, 692, 758Gofford J., Reeves J. N., Tombesi F., Braito V., Turner T. J.,Miller L., Cappi M., 2013, MNRAS, 430, 60 Gofford J., et al., 2014, ApJ, 784, 77Gofford J., Reeves J. N., McLaughlin D. E., Braito V., TurnerT. J., Tombesi F., Cappi M., 2015, MNRAS, 451, 4169Grier C. J., et al., 2015, ApJ, 806, 111Hagino K., Odaka H., Done C., Gandhi P., Watanabe S., SakoM., Takahashi T., 2015, MNRAS, 446, 663Hamann F., 1997, ApJS, 109, 279Hamann F., 1998, ApJ, 500, 798Hamann F., Ferland G., 1999, ARA&A, 37, 487Hamann F., Sabra B., 2004, in Richards G. T., Hall P. B., eds,Astronomical Society of the Pacific Conference Series Vol. 311,AGN Physics with the Sloan Digital Sky Survey. p. 203Hamann F., Barlow T., Cohen R. D., Junkkarinen V., BurbidgeE. M., 1997a, in Arav N., Shlosman I., Weymann R. J., eds,Astronomical Society of the Pacific Conference Series Vol. 128,Mass Ejection from Active Galactic Nuclei. p. 19Hamann F., Barlow T. A., Junkkarinen V., Burbidge E. M.,1997b, ApJ, 478, 80Hamann F., Barlow T. A., Junkkarinen V., 1997c, ApJ, 478, 87Hamann F., Kanekar N., Prochaska J. X., Murphy M. T., EllisonS., Malec A. L., Milutinovic N., Ubachs W., 2011, MNRAS,410, 1957Hamann F., Chartas G., McGraw S., Rodriguez Hidalgo P.,Shields J., Capellupo D., Charlton J., Eracleous M., 2013,MNRAS, 435, 133Hamann F., et al., 2017, MNRAS, 464, 3431Higginbottom N., Proga D., Knigge C., Long K. S., MatthewsJ. H., Sim S. A., 2014, ApJ, 789, 19Hopkins P. F., Elvis M., 2010, MNRAS, 401, 7Hopkins P. F., Hernquist L., Cox T. J., Kereˇs D., 2008, ApJS,175, 356Jannuzi B. T., et al., 1996, ApJ, 470, L11+Kaspi S., et al., 2002, ApJ, 574, 643Kaspi S., Maoz D., Netzer H., Peterson B. M., Vestergaard M.,Jannuzi B. T., 2005, ApJ, 629, 61Knigge C., Scaringi S., Goad M. R., Cottis C. E., 2008, MNRAS,386, 1426Korista K. T., Voit G. M., Morris S. L., Weymann R. J., 1993,ApJS, 88, 357Laor A., Davis S. W., 2014, MNRAS, 438, 3024Leighly K. M., 2004, ApJ, 611, 125Leighly K. M., Halpern J. P., Jenkins E. B., Casebeer D., 2007,ApJS, 173, 1Leighly K. M., Hamann F., Casebeer D. A., Grupe D., 2009, ApJ,701, 176Leighly K. M., Dietrich M., Barber S., 2011, ApJ, 728, 94Leighly K. M., Cooper E., Grupe D., Terndrup D. M., KomossaS., 2015, ApJ, 809, L13Luo B., et al., 2015, ApJ, 805, 122Matzeu G. A., Reeves J. N., Nardini E., Braito V., Costa M. T.,Tombesi F., Gofford J., 2016, MNRAS, 458, 1311Matzeu G. A., Reeves J. N., Braito V., Nardini E., McLaughlinD. E., Lobban A. P., Tombesi F., Costa M. T., 2017, preprint,( arXiv:1708.03546 )Misawa T., Eracleous M., Charlton J. C., Kashikawa N., 2007,ApJ, 660, 152Misawa T., Charlton J. C., Eracleous M., 2014a, ApJ, 792, 77Misawa T., Inada N., Oguri M., Gandhi P., Horiuchi T., Koya-mada S., Okamoto R., 2014b, ApJ, 794, L20Moravec E. A., Hamann F., Capellupo D. M., McGraw S. M.,Shields J. C., Rodr´ıguez Hidalgo P., 2017, MNRAS, 468, 4539Nagao T., Marconi A., Maiolino R., 2006, A&A, 447, 157Nardini E., et al., 2015, Science, 347, 860Netzer H., et al., 2003, ApJ, 599, 933O’Brien P. T., Reeves J. N., Simpson C., Ward M. J., 2005, MN-RAS, 360, L25Pˆaris I., Petitjean P., Ross N. P., Myers A. D., Aubourg ´E., Stre-blyanska A., Bailey S. e. a., 2017, A&A, 597, A79MNRAS , 1–11 (2015) F. Hamann et al.
Reeves J. N., O’Brien P. T., Ward M. J., 2003, ApJ, 593, L65Reeves J. N., et al., 2009, ApJ, 701, 493Reeves J. N., et al., 2014, ApJ, 780, 45Reeves J. N., Braito V., Nardini E., Behar E., O’Brien P. T.,Tombesi F., Turner T. J., Costa M. T., 2016, ApJ, 824, 20Richards G. T., et al., 2011, AJ, 141, 167Rodr´ıguez Hidalgo P., 2008, PhD thesis, University ofFlorida, Gainesville, http://etd.fcla.edu/UF/UFE0022674/rodriguez_p.pdf
Rodr´ıguez Hidalgo P., Hamann F., Hall P., 2011, MNRAS, 411,247Rodr´ıguez Hidalgo P., Eracleous M., Charlton J., Hamann F.,Murphy M. T., Nestor D., 2013, ApJ, 775, 14Rogerson J. A., Hall P. B., Rodr´ıguez Hidalgo P., Pirkola P.,Brandt W. N., Filiz Ak N., 2016, MNRAS, 457, 405Ross N. P., Myers A. D., Sheldon E. S., Y`eche C., Strauss M. A.,Bovy J., Kirkpatrick J. A. e. a., 2012, ApJS, 199, 3Rupke D. S. N., Veilleux S., 2013, ApJ, 768, 75Sabra B. M., Hamann F., Jannuzi B. T., George I. M., ShieldsJ. C., 2003, ApJ, 590, 66Saez C., Chartas G., 2011, ApJ, 737, 91Sim S. A., Proga D., Miller L., Long K. S., Turner T. J., 2010,MNRAS, 408, 1396Simpson C., Ward M., O’Brien P., Reeves J., 1999, MNRAS, 303,L23Steffen A. T., Strateva I., Brandt W. N., Alexander D. M., Koeke-moer A. M., Lehmer B. D., Schneider D. P., Vignali C., 2006,AJ, 131, 2826Steidel C. C., 1990, ApJS, 74, 37Stern J., Behar E., Laor A., Baskin A., Holczer T., 2014, MNRAS,445, 3011Strateva I. V., Brandt W. N., Schneider D. P., Vanden Berk D. G.,Vignali C., 2005, AJ, 130, 387Takeuchi S., Ohsuga K., Mineshige S., 2013, PASJ, 65, 88Tarter C. B., Tucker W. H., Salpeter E. E., 1969, ApJ, 156, 943Telfer R. C., Kriss G. A., Zheng W., Davidsen A. F., Green R. F.,1998, ApJ, 509, 132Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., YaqoobT., Braito V., Dadina M., 2010, A&A, 521, A57Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., BraitoV., Dadina M., 2011, ApJ, 742, 44Tombesi F., Cappi M., Reeves J. N., Braito V., 2012, MNRAS,422, L1Tombesi F., Cappi M., Reeves J. N., Nemmen R. S., Braito V.,Gaspari M., Reynolds C. S., 2013, MNRAS, 430, 1102Torres C. A. O., Quast G. R., Coziol R., Jablonski F., de la RezaR., L´epine J. R. D., Greg´orio-Hetem J., 1997, ApJ, 488, L19Trump J. R., et al., 2006, ApJS, 165, 1Warner C., Hamann F., Dietrich M., 2004, ApJ, 608, 136Waters T., Proga D., Dannen R., Kallman T. R., 2017, MNRAS,467, 3160Weymann R. J., Morris S. L., Foltz C. B., Hewett P. C., 1991,ApJ, 373, 23Wu J., et al., 2011, ApJ, 736, 28Wu J., Brandt W. N., Anderson S. F., Diamond-Stanic A. M.,Hall P. B., Plotkin R. M., Schneider D. P., Shemmer O., 2012,ApJ, 747, 10Zakamska N. L., et al., 2016, MNRAS, 459, 3144de Kool M., Korista K. T., Arav N., 2002, ApJ, 580, 54This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000