Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Constraining the geometry of AGN outflows with reflection spectroscopy

Constraining the geometry of AGN outflows with reflection spectroscopy Constraining out ow geometries 1 Constraining the geometry of AGN out ows with re ection spectroscopy 1? 2 2 3 4 M. L. Parker, D. J. K. Buisson, J. Jiang (Ü H), L. C. Gallo, E. Kara, 1 2 G. A. Matzeu and D. J. Walton European Space Agency (ESA), European Space Astronomy Centre (ESAC), E-28691 Villanueva de la Can~ada, Madrid, Spain Institute of Astronomy, University of Cambridge, Madingley Road, CB3 0HA Cambridge, UK Saint Mary's University, Department of Astronomy & Physics, 923 Robie Street, Halifax, Canada, B3H 3C3 Department of Astronomy, University of Maryland, College Park, MD 20742, USA Accepted XXX. Received YYY; in original form ZZZ ABSTRACT We collate active galactic nuclei (AGN) with reported detections of both relativistic re ection and ultra-fast out ows. By comparing the inclination of the inner disc from re ection with the line-of- sight velocity of the out ow, we show that it is possible to meaningfully constrain the geometry of the absorbing material. We nd a clear relation between the velocity and inclination, and demonstrate that it can potentially be explained either by simple wind geometries or by absorption from the disc surface. Due to systematic errors and a shortage of high-quality simultaneous measurements our conclusions are tentative, but this study represents a proof-of-concept that has great potential. Key words: accretion, accretion discs { galaxies: active { black hole physics 1 INTRODUCTION or radiation pressure (Reeves et al. 2003; Fukumura et al. 2015), and are one possibility for driving AGN feedback There are two main forms of relativistic spectroscopy in X- (e.g. Fabian 2012). An alternative possibility for produc- ray studies of AGN: relativistic re ection, and ultra-fast out- ing these features is in an absorbing layer on the surface ows. Both rely on detecting red and blue-shifted elemental of the disc (Gallo & Fabian 2011, 2013), where the extreme emission or absorption lines, usually from iron, and both velocity arises from the orbital motion of the gas and the reach velocities of a signi cant fraction of c. These two tech- absorption is imprinted on the re ection spectrum. Because niques are rarely used together, so the opportunities a orded we only have one line of sight (LOS) through the gas, it is by combining the two remain largely unexplored. extremely hard to directly measure the density and hence Relativistic re ection occurs when relatively cool, dense location of the absorbing material, so we must use other material in the accretion disc around a black hole is illumi- approaches to try and constrain the geometry of the gas. nated by X-rays. A featureless X-ray continuum is repro- While there is no reason to expect viewing angle to be cessed into a series of uorescent emission lines, absorption the sole determiner of the line of sight velocity, it should edges, and a Compton scattered hump (George & Fabian certainly have an impact, which depends on the launch- 1991). This characteristic re ection spectrum is then blurred ing mechanism. Radiation pressure driven disc winds are and shifted by a combination of Doppler shifting, special rel- expected to have an equatorial geometry (e.g. Proga et al. ativistic boosting, and gravitational redshift (Fabian et al. 2000), so velocity would generally increase with inclination, 1989). By measuring the extent of these relativistic e ects on whereas MHD wind simulations predict higher velocities at the pro le of the Fe K line (or other lines, e.g. Fabian et al. low inclinations (Fukumura et al. 2010). In this work, we ex- 2009; Madej et al. 2014), we can infer physical parameters amine the possibility of constraining the geometry of UFOs of the black hole and accretion disc, such as the inclination using their inclination dependence, taking the inclination of the disc and the black hole spin parameter (see review by values from re ection modeling. Reynolds 2014). Ultra-fast out ows (UFOs) are identi ed by strongly blueshifted absorption lines in X-ray spectra (Chartas et al. 2 SAMPLE 2002; Pounds et al. 2003). They are generally thought to be due to winds from the AGN disc, accelerated by magnetic We performed a literature search to identify sources which have both absorption from an out ow with a well constrained velocity and a well constrained inclination E-mail: [email protected] from relativistic re ection spectroscopy. In a few cases Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 2 M. L. Parker et al. (Tombesi et al. 2011a; Parker et al. 2017, Parker et al., sub- and physically plausible. For the sake of keeping as large a mitted) both are measured simultaneously, but for the ma- sample as possible, we do not implement these as a strict re- jority these measurements are taken from di erent papers. quirement. Rather, we select all sources with a constrained In general, we prioritise results from papers presenting spec- inclination and ag those which do not meet these criteria. tral tting of an individual source over those where a large Then, in Section 3, we run the analysis on the full sample sample of sources are analysed. We also prefer more re- and on those results that meet the quality control criteria. cent papers, as these are likely to have higher quality Where the errors on the inclination are less than 5, datasets available, as well as the latest models. The selec- we make the conservative assumption that the true un- tion of UFO and re ection results is presented in detail certainty is dominated by systematic errors of 5(see e.g. below, and the values are given in Table 1. The data and the di erence between relxill and reflionx discussed in code used for the analysis in Section 3 are available here: Middleton et al. 2016). This applies in most cases. Finally, https://github.com/M-L-Parker/UFO_inclinations. we exclude sources that require retrograde spin or signi cant disk truncation. In this case it is likely that the line is dom- inated by a narrow core, so the parameters are not reliable. 2.1 UFOs The observed inclinations are strongly concentrated at 40{ 50, which is most likely due to selection e ects. At higher To avoid contamination from warm absorbers, we implement inclinations, the LOS is likely to intersect the torus, ob- a velocity cut-o at 0.033c (10,000 km s ; Tombesi et al. scuring the nucleus, and at lower inclinations the relativistic 2010).These may or may not be related to UFOs (see dis- blurring is weaker and correspondingly harder to measure. cussion in e.g. Tombesi et al. 2013; Pounds & King 2013; Laha et al. 2016), and in general have small velocities or are consistent with the source rest-frame (Laha et al. 2014). The treatment of multiple detections must be care- 3 RESULTS fully considered. We wish to avoid having the sample dom- 3.1 Correlation analysis inated by a small number of sources with multiple de- tections of the same out ow at di erent velocities. For We show the UFO velocity against re ection inclination example, the UFO in PDS 456 appears at slightly dif- points in Figs 1 and 3, overplotted with simple models (see ferent velocities (from  0:23{0.33c) in almost every ob- Section 3.2). We use a Monte-Carlo approach to estimate the servation taken of the source (Go ord et al. 2014), most signi cance of any correlation between the two parameters. likely due to a ux-dependent velocity shift (Matzeu et al. The distributions of v and i are approximately log-normal 2017).However, an additional layer of ultra-fast absorption and normal, respectively, so we randomly draw 100,000 sets is present in PDS 456 (Reeves et al. 2018b) simultaneously, of points from distributions with the same mean and stan- which should be included separately. This is further compli- dard deviation (Mean log(v)= 0:89,  = 0:29, mean log(v) cated by the transient behaviour of out owing absorption i = 44:4,  = 9:53). Of these simulated sets of points, we lines (e.g. Cappi et al. 2009), which is likely due to the gas nd 37 that exceed the Pearson correlation coecient of the being fully ionized at high source uxes (Pinto et al. 2018) real data (0.64) and 629 that exceed the Spearman correla- and inhomogeneities in the wind. To mitigate this, we com- tion coecient (0.52). This gives probabilities of 0.0004 and bine multiple UFO detections from a single source into an 0.006 for a correlation this strong occurring from randomly averaged value when the velocity estimates are within the distributed points. Excluding the two sources where the re- reported errors of each other, within 0.01c of each other, or ection modeling does not meet the Reynolds (2014) crite- within 10% of each other, whichever is largest. Where the ria marginally strengthens the correlation (Pearson r =0.63, same authors have written multiple papers on a particular P = 0:0011, Spearman r = 0:49, P = 0:017). source, taking multiple velocity measurements, we assume that the later papers supersede earlier ones, and only use the latest values unless the out ows presented are clearly 3.2 Models distinct. Otherwise, we include all values in our analysis. Regardless of whether a linear correlation is a signi cant im- provement over the null hypothesis of randomly distributed points, it is still possible to use these points to infer some- 2.2 Relativistic re ection thing about the geometry of UFOs. We construct a simple Obtaining a corresponding set of re ection measurements imitation of a stream-line, where a thin out ow starts mov- is simpler than the UFO case, as there should only be one ing vertically from the disc at some radius r , following a launch value of the inclination for each AGN. We take the latest circular path with radius r . For this purpose, the units of curve available measurement for each source, unless there is an radius are arbitrary. Once it reaches a nal inclination i final earlier NuSTAR paper, in which case we use that value (in it leaves the circular path and travels on the tangent with general, NuSTAR's high energy coverage gives a better con- i = i . We also assume that the UFO accelerates along final straint on the re ection spectrum than soft instruments). this path, following v = v (1 R =(R + r )) (adapted from inf v v Reynolds (2014) gives a list of `quality control' crite- Knigge et al. 1995; Sim et al. 2008), where is a constant ria for AGN spin measurements based on re ection spec- (and = 0 gives a constant velocity) and R is a charac- troscopy. In brief, Reynolds suggests that: a full ionized re- teristic length scale. We assume that the X-ray source is ection model must be used; the iron abundance must be a coincident with the black hole (h = 0, r = 0), and ignore all free parameter; the inclination parameter must be free, and relativistic e ects. This geometry is shown in Fig. 2. This ge- constrained; and the emissivity index must be free to vary ometry is intended as an approximation of that in radiation- Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 Constraining out ow geometries 3 Table 1. Out ow velocities and inclinations for the sources in our sample. Velocities with multiple references are the result of taking the weighted average of multiple measurements that meet the criteria discussed in Section 2.1. Similarly, sources with multiple velocities are those where multiple measurements did not meet the criteria for merging. Name v (c) Reference i (degrees) Reference UFO +7:0 1H 0419-577 0:079  0:007 Tombesi et al. (2011b) 49:0 Walton et al. (2013) 4:0 z +0:01 1H 0707-495 0:11 Dauser et al. (2012) 43:0  2:0 Kara et al. (2015) 0:02 0:18  0:01 Hagino et al. (2016); Dauser et al. (2012) z y 3C 111 0:105  0:006 Go ord et al. (2013); Tombesi et al. (2011a) 44:0  2:0 Tombesi et al. (2011a) +5:0 Ark 120 0:29  0:02 Tombesi et al. (2011b) 45:0 Garc a et al. (2014) 2:0 IC 4329A 0:098  0:004 Tombesi et al. (2011b) 35:0  5:0 Mantovani et al. (2014) z +0:007 +3 IRAS 00521-7054 0:403 Walton et al. (in prep) 63 Walton et al. (in prep) 0:006 2 IRAS 13224-3809 0:236  0:006 Parker et al. (2017) 59:0  1:0 Parker et al. (2017) +2:0 IRAS 13349+2438 0:13  0:01 Parker et al. (submitted) 48:0 Parker et al. (submitted) 1:0 MCG-5-23-16 0:116  0:004 Tombesi et al. (2011b) 51:0  7:0 Zoghbi et al. (2017) +3:0 MR 2251-178 0:137  0:008 Go ord et al. (2013) 24:0 Nardini et al. (2014) 5:0 Mrk 1044 0:10  0:01 Mallick et al. (submitted) 47  3 Mallick et al. (submitted) +5:0 Mrk 509 0:14  0:0024 Cappi et al. (2009); Tombesi et al. (2011b) 50:0 Walton et al. (2013) 3:0 0:171  0:003 Cappi et al. (2009); Tombesi et al. (2011b) 0:197  0:005 Cappi et al. (2009); Tombesi et al. (2011b) +6:0 Mrk 766 0:039  0:03 Go ord et al. (2013) 39:0 Buisson et al. (submitted) 3:0 0:082  0:006 Tombesi et al. (2011b) Mrk 79 0:092  0:004 Tombesi et al. (2011b) 24:0  1:0 Gallo et al. (2011) +6:0 Mrk 841 0:055  0:025 Tombesi et al. (2011b) 46:0 Walton et al. (2013) 5:0 NGC 4051 0:202  0:006 Tombesi et al. (2011b) 37:0  5:0 Risaliti et al. (in prep) NGC 4151 0:0452  0:0099 Go ord et al. (2013); Patrick et al. (2012) < 10 Beuchert et al. (2017) 0:106  0:007 Tombesi et al. (2011b) +0:1 NGC 5506 0:246  0:006 Go ord et al. (2013) 41:0 Sun et al. (2017) 0:2 PDS 456 0:278  0:003 Reeves et al. (2018b); Matzeu et al. (2017) 65:0  2:0 Chiang et al. (2017) 0:46  0:02 Reeves et al. (2018b) PG 1211+143 0:0598  0:00069 Pounds et al. (2016); Danehkar et al. (2018); 44:0  2:0 Lobban et al. (2016) Reeves et al. (2018a) 0:129  0:002 Pounds et al. (2016) 0:151  0:003 Tombesi et al. (2011b) Swift J2127 0:231  0:006 Go ord et al. (2013) 49:0  2:0 Marinucci et al. (2014) These results do not meet the quality-control criteria of Reynolds (2014). These sources have joint re ection/UFO tting, which either gives the result reported here or is consistent with it. driven winds (e.g. Proga et al. 2000), as these give a simple locity of the absorption line and the viewing angle is to pro- explanation for the higher velocities at higher inclinations. duce the absorption in the disc itself. This model, explored MHD winds predict concave stream lines (Fukumura et al. by Gallo & Fabian (2011, 2013), explains UFO absorption 2010), so give higher velocities at small inclinations. How- using a surface layer on the disc, with the strong blueshift ever, the exact pattern observed depends on the ionization due to the orbital velocity of the absorbing material, rather and density structure within the wind, so it may still be than an out owing wind. In this case, the inclination depen- possible to explain these results in an MHD scenario. dence of the absorption velocity arises from the increased LOS velocity of the disc at high inclinations. As a simple In most cases where the line of sight intersects the wind proxy for the velocity of an absorption line from the surface in this model, it crosses the wind twice. Once while the wind of the disc, we take the maximum blueshift from the relline is rising steeply, and once in the tail where the gradient is model (Dauser et al. 2010). This gives a simple correlation constant. Because of the radial acceleration assumed and between i and v (shown in Fig. 3), although it does not reach close alignment with the line of sight in the tail, this inter- high enough velocities to account for the most blueshifted section results in a much higher apparent velocity. A simple absorption lines. The only parameter of this model is the way of only producing one measurable value for the velocity black hole spin (a), but this has a limited e ect as the max- is to assume that the gas where the rst intersection with the imum blueshift of the disc comes from further out than the LOS occurs is fully ionized. In this case, no absorption lines innermost stable circular orbit. We note that this predicted would be produced, and only the second intersection would velocity should be an upper limit as it is the maximum found be observed. An example that provides a reasonable match on the disc, so points should generally lie below the line. to the data is shown in Fig. 1, with parameters r = 10, launch r = 300, v = 0:5c, and R = 1000. We show the e ect curve v inf of varying the acceleration coecient and nal inclination 4 DISCUSSION i . From this it is clear that meaningful constraints on final these parameters can be obtained. We note that there are other possible explanations for broad An alternative way of giving a relation between the ve- emission lines in AGN. For example, Nardini et al. (2015) Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 r curve LOS final 4 M. L. Parker et al. 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 inclination (degrees) inclination (degrees) Figure 1. UFO velocity as a function of re ection inclination, overplotted with a toy model for an out owing disc wind. Left: the e ect of changing the acceleration coecient to 0.5 (red,left), 1.0 (blue, middle) and 1.5 (green, right) with nal inclination i = 65 . final Right: the e ect of setting i to 45 (red, left), 65 (blue, middle) and 85 (green, right) with = 1. final 500 interpret the emission line in PDS 456 as a P-Cygni pro- le, where the Fe K emission is produced by scattering o the out owing wind. In this case, the relativistic broad- ening is produced by the velocity of the wind instead of the orbits in the disc (Done et al. 2007). This model relies on partial-covering absorption to explain most of the spec- tral complexity. However, Chiang et al. (2017) present an alternative model where the broad-band spectrum can be fully explained by relativistic re ection, warm absorption, launch and the UFO. This interpretation is supported by the de- tection of a soft X-ray lag, generally interpreted as rever- beration in the inner disc and found in many unobscured 0 100 200 300 400 500 600 700 sources (De Marco et al. 2013). While the detection of an X-ray lag is usually considered strong evidence for the pres- ence of relativistic re ection in a source (and many of the Figure 2. Simple streamline geometry, with the wind shown in sources in our sample show reverberation lags, Kara et al. blue and the line of sight in red (dashed). Length units are arbi- 2016; De Marco et al. 2013), it is dicult to rule out a con- trary. tribution to the total broad-line pro le from scattering in the wind. Indeed, given that disc winds are likely to be most 0.5 dense at the point they launch from the disc and should be a = 0.998 co-rotating with it, then scattering from the disc surface and a = 0 wind may be thought of as a single continuous process. The 0.4 impact of having Fe K emission from both the disc and wind is not understood, but could have a signi cant e ect on the measured inclination. A higher velocity out ow will produce 0.3 more blueshifted emission, in the same way that a higher in- clination gives a more blueshifted line pro le for disc re ec- tion. We note that the tentative relation identi ed here may 0.2 be indicative of the e ect of wind emission on the net rel- ativistic line pro le rather than an inclination dependence. 0.1 A related issue is the lack of simultaneous UFO/re ection modeling. Using a simpli ed phenomenological continuum may exaggerate the signi cance of line features, leading to 0.0 false detections (Zoghbi et al. 2015). Similarly, not account- 0 10 20 30 40 50 60 70 80 inclination (degrees) ing for UFO absorption during re ection modeling may bias the measured parameters. We will revisit the spectra of some Figure 3. Toy model for absorption from a layer on the disc. of these sources for joint modeling to investigate this further in future work. We have assumed throughout that all UFOs have the Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 v/c v/c v/c Constraining out ow geometries 5 same shape and velocity pro le, and that the observed ve- Fukumura K., Tombesi F., Kazanas D., Shrader C., Behar E., Contopoulos I., 2015, ApJ, 805, 17 locity is solely determined by the LOS angle. This is unlikely Gallo L. C., Fabian A. C., 2011, MNRAS, 418, L59 to be the case in practice: the accretion rate, for example, is Gallo L. C., Fabian A. C., 2013, MNRAS, 434, L66 likely to have a major impact on the velocity of the material. Gallo L. C., Miniutti G., Miller J. M., Brenneman L. W., Fabian Similarly, we have implicitly assumed that wind instabilities A. C., Guainazzi M., Reynolds C. S., 2011, MNRAS, 411, 607 play a negligible role in determining the observed velocity, Garc a J., et al., 2014, ApJ, 782, 76 which is unlikely. Another caveat is that our sample is any- George I. M., Fabian A. C., 1991, MNRAS, 249, 352 thing but unbiased, and the biases are poorly understood. Go ord J., Reeves J. N., Tombesi F., Braito V., Turner T. J., Re ection measurements are generally biased towards high Miller L., Cappi M., 2013, MNRAS, 430, 60 spin (Vasudevan et al. 2016), and there may be a similar Go ord J., et al., 2014, ApJ, 784, 77 bias towards high inclination, as it produces broader, easier Hagino K., Odaka H., Done C., Tomaru R., Watanabe S., Taka- hashi T., 2016, MNRAS, 461, 3954 to measure lines. Similarly, it is plausible that UFO veloc- Kara E., et al., 2015, MNRAS, 449, 234 ity measurements are biased towards lower velocities, as the Kara E., Alston W. N., Fabian A. C., Cackett E. M., Uttley P., sensitivity of X-ray detectors typically declines with energy Reynolds C. S., Zoghbi A., 2016, MNRAS, 462, 511 (although this may be remedied as the NuSTAR archive Knigge C., Woods J. A., Drew J. E., 1995, MNRAS, 273, 225 grows). Laha S., Guainazzi M., Dewangan G. C., Chakravorty S., Kemb- havi A. K., 2014, MNRAS, 441, 2613 Laha S., Guainazzi M., Chakravorty S., Dewangan G. C., Kemb- 5 CONCLUSIONS havi A. K., 2016, MNRAS, 457, 3896 Lobban A. P., Pounds K., Vaughan S., Reeves J. N., 2016, ApJ, We have identi ed a correlation between the velocity of 831, 201 highly ionized absorption features from UFOs and the in- Madej O. K., Garc a J., Jonker P. G., Parker M. L., Ross R., clination of the inner accretion disc measured from re ec- Fabian A. C., Chenevez J., 2014, MNRAS, 442, 1157 tion spectroscopy. The correlation is formally signi cant, but Mantovani G., Nandra K., Ponti G., 2014, MNRAS, 442, L95 Marinucci A., et al., 2014, MNRAS, 440, 2347 heavily reliant on a small number of points at high velocity. Matzeu G. A., Reeves J. N., Braito V., Nardini E., McLaughlin We show that the observed points can be explained by D. E., Lobban A. P., Tombesi F., Costa M. T., 2017, MNRAS, simple toy models of an out owing wind or absorption from 472, L15 a disc, although the latter cannot account for the highest Middleton M. J., Parker M. L., Reynolds C. S., Fabian A. C., velocity features. With more detailed modeling and higher Loh nk A. M., 2016, MNRAS, 457, 1568 quality data, this technique could be very powerful for con- Nardini E., Reeves J. N., Porquet D., Braito V., Grosso N., Gof- straining the geometry of out owing material in AGN. ford J., 2014, MNRAS, 440, 1200 Nardini E., et al., 2015, Science, 347, 860 Parker M. L., et al., 2017, Nature, 543, 83 Patrick A. R., Reeves J. N., Porquet D., Markowitz A. G., Braito ACKNOWLEDGEMENTS V., Lobban A. P., 2012, MNRAS, 426, 2522 MLP and GM are supported by European Space Agency Pinto C., et al., 2018, MNRAS, 476, 1021 (ESA) Research Fellowships. JJ acknowledges support by Pounds K. A., King A. R., 2013, MNRAS, 433, 1369 the Cambridge Trust and the Chinese Scholarship Coun- Pounds K. A., Reeves J. N., Page K. L., Wynn G. A., O'Brien P. T., 2003, MNRAS, 342, 1147 cil Joint Scholarship Programme (201604100032). DJKB is Pounds K., Lobban A., Reeves J., Vaughan S., 2016, MNRAS, supported by an STFC studentship. 457, 2951 Proga D., Stone J. M., Kallman T. R., 2000, ApJ, 543, 686 Reeves J. N., O'Brien P. T., Ward M. J., 2003, ApJ, 593, L65 REFERENCES Reeves J. N., Lobban A., Pounds K. A., 2018a, ApJ, 854, 28 Reeves J. N., Braito V., Nardini E., Lobban A. P., Matzeu G. A., Beuchert T., et al., 2017, A&A, 603, A50 Costa M. T., 2018b, ApJ, 854, L8 Cappi M., et al., 2009, A&A, 504, 401 Reynolds C. S., 2014, Space Sci. Rev., 183, 277 Chartas G., Brandt W. N., Gallagher S. C., Garmire G. P., 2002, Sim S. A., Long K. S., Miller L., Turner T. J., 2008, MNRAS, ApJ, 579, 169 388, 611 Chiang C.-Y., Cackett E. M., Zoghbi A., Fabian A. C., Kara E., Sun S., Guainazzi M., Ni Q., Wang J., Qian C., Shi F., Wang Y., Parker M. L., Reynolds C. S., Walton D. J., 2017, MNRAS, Bambi C., 2017, preprint, (arXiv:1704.03716) 472, 1473 Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Yaqoob Danehkar A., et al., 2018, ApJ, 853, 165 T., Braito V., Dadina M., 2010, A&A, 521, A57 Dauser T., Wilms J., Reynolds C. S., Brenneman L. W., 2010, Tombesi F., Sambruna R. M., Reeves J. N., Reynolds C. S., Braito MNRAS, 409, 1534 V., 2011a, MNRAS, 418, L89 Dauser T., et al., 2012, MNRAS, 422, 1914 Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Braito De Marco B., Ponti G., Cappi M., Dadina M., Uttley P., Cackett V., Dadina M., 2011b, ApJ, 742, 44 E. M., Fabian A. C., Miniutti G., 2013, MNRAS, 431, 2441 Tombesi F., Cappi M., Reeves J. N., Nemmen R. S., Braito V., Done C., Sobolewska M. A., Gierlinsk  i M., Schurch N. J., 2007, Gaspari M., Reynolds C. S., 2013, MNRAS, 430, 1102 MNRAS, 374, L15 Vasudevan R. V., Fabian A. C., Reynolds C. S., Aird J., Dauser Fabian A. C., 2012, ARA&A, 50, 455 T., Gallo L. C., 2016, MNRAS, 458, 2012 Fabian A. C., Rees M. J., Stella L., White N. E., 1989, MNRAS, Walton D. J., Nardini E., Fabian A. C., Gallo L. C., Reis R. C., 238, 729 2013, MNRAS, 428, 2901 Fabian A. C., et al., 2009, Nature, 459, 540 Zoghbi A., et al., 2015, ApJ, 799, L24 Fukumura K., Kazanas D., Contopoulos I., Behar E., 2010, ApJ, Zoghbi A., et al., 2017, ApJ, 836, 2 715, 636 Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Monthly Notices of the Royal Astronomical Society Letters Oxford University Press

Constraining the geometry of AGN outflows with reflection spectroscopy

Loading next page...
 
/lp/ou_press/constraining-the-geometry-of-agn-outflows-with-reflection-spectroscopy-SIDEJ80zwE

References (1)

Publisher
Oxford University Press
Copyright
© 2018 The Author(s) Published by Oxford University Press on behalf of the Royal Astronomical Society
ISSN
1745-3925
eISSN
1745-3933
DOI
10.1093/mnrasl/sly096
Publisher site
See Article on Publisher Site

Abstract

Constraining out ow geometries 1 Constraining the geometry of AGN out ows with re ection spectroscopy 1? 2 2 3 4 M. L. Parker, D. J. K. Buisson, J. Jiang (Ü H), L. C. Gallo, E. Kara, 1 2 G. A. Matzeu and D. J. Walton European Space Agency (ESA), European Space Astronomy Centre (ESAC), E-28691 Villanueva de la Can~ada, Madrid, Spain Institute of Astronomy, University of Cambridge, Madingley Road, CB3 0HA Cambridge, UK Saint Mary's University, Department of Astronomy & Physics, 923 Robie Street, Halifax, Canada, B3H 3C3 Department of Astronomy, University of Maryland, College Park, MD 20742, USA Accepted XXX. Received YYY; in original form ZZZ ABSTRACT We collate active galactic nuclei (AGN) with reported detections of both relativistic re ection and ultra-fast out ows. By comparing the inclination of the inner disc from re ection with the line-of- sight velocity of the out ow, we show that it is possible to meaningfully constrain the geometry of the absorbing material. We nd a clear relation between the velocity and inclination, and demonstrate that it can potentially be explained either by simple wind geometries or by absorption from the disc surface. Due to systematic errors and a shortage of high-quality simultaneous measurements our conclusions are tentative, but this study represents a proof-of-concept that has great potential. Key words: accretion, accretion discs { galaxies: active { black hole physics 1 INTRODUCTION or radiation pressure (Reeves et al. 2003; Fukumura et al. 2015), and are one possibility for driving AGN feedback There are two main forms of relativistic spectroscopy in X- (e.g. Fabian 2012). An alternative possibility for produc- ray studies of AGN: relativistic re ection, and ultra-fast out- ing these features is in an absorbing layer on the surface ows. Both rely on detecting red and blue-shifted elemental of the disc (Gallo & Fabian 2011, 2013), where the extreme emission or absorption lines, usually from iron, and both velocity arises from the orbital motion of the gas and the reach velocities of a signi cant fraction of c. These two tech- absorption is imprinted on the re ection spectrum. Because niques are rarely used together, so the opportunities a orded we only have one line of sight (LOS) through the gas, it is by combining the two remain largely unexplored. extremely hard to directly measure the density and hence Relativistic re ection occurs when relatively cool, dense location of the absorbing material, so we must use other material in the accretion disc around a black hole is illumi- approaches to try and constrain the geometry of the gas. nated by X-rays. A featureless X-ray continuum is repro- While there is no reason to expect viewing angle to be cessed into a series of uorescent emission lines, absorption the sole determiner of the line of sight velocity, it should edges, and a Compton scattered hump (George & Fabian certainly have an impact, which depends on the launch- 1991). This characteristic re ection spectrum is then blurred ing mechanism. Radiation pressure driven disc winds are and shifted by a combination of Doppler shifting, special rel- expected to have an equatorial geometry (e.g. Proga et al. ativistic boosting, and gravitational redshift (Fabian et al. 2000), so velocity would generally increase with inclination, 1989). By measuring the extent of these relativistic e ects on whereas MHD wind simulations predict higher velocities at the pro le of the Fe K line (or other lines, e.g. Fabian et al. low inclinations (Fukumura et al. 2010). In this work, we ex- 2009; Madej et al. 2014), we can infer physical parameters amine the possibility of constraining the geometry of UFOs of the black hole and accretion disc, such as the inclination using their inclination dependence, taking the inclination of the disc and the black hole spin parameter (see review by values from re ection modeling. Reynolds 2014). Ultra-fast out ows (UFOs) are identi ed by strongly blueshifted absorption lines in X-ray spectra (Chartas et al. 2 SAMPLE 2002; Pounds et al. 2003). They are generally thought to be due to winds from the AGN disc, accelerated by magnetic We performed a literature search to identify sources which have both absorption from an out ow with a well constrained velocity and a well constrained inclination E-mail: [email protected] from relativistic re ection spectroscopy. In a few cases Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 2 M. L. Parker et al. (Tombesi et al. 2011a; Parker et al. 2017, Parker et al., sub- and physically plausible. For the sake of keeping as large a mitted) both are measured simultaneously, but for the ma- sample as possible, we do not implement these as a strict re- jority these measurements are taken from di erent papers. quirement. Rather, we select all sources with a constrained In general, we prioritise results from papers presenting spec- inclination and ag those which do not meet these criteria. tral tting of an individual source over those where a large Then, in Section 3, we run the analysis on the full sample sample of sources are analysed. We also prefer more re- and on those results that meet the quality control criteria. cent papers, as these are likely to have higher quality Where the errors on the inclination are less than 5, datasets available, as well as the latest models. The selec- we make the conservative assumption that the true un- tion of UFO and re ection results is presented in detail certainty is dominated by systematic errors of 5(see e.g. below, and the values are given in Table 1. The data and the di erence between relxill and reflionx discussed in code used for the analysis in Section 3 are available here: Middleton et al. 2016). This applies in most cases. Finally, https://github.com/M-L-Parker/UFO_inclinations. we exclude sources that require retrograde spin or signi cant disk truncation. In this case it is likely that the line is dom- inated by a narrow core, so the parameters are not reliable. 2.1 UFOs The observed inclinations are strongly concentrated at 40{ 50, which is most likely due to selection e ects. At higher To avoid contamination from warm absorbers, we implement inclinations, the LOS is likely to intersect the torus, ob- a velocity cut-o at 0.033c (10,000 km s ; Tombesi et al. scuring the nucleus, and at lower inclinations the relativistic 2010).These may or may not be related to UFOs (see dis- blurring is weaker and correspondingly harder to measure. cussion in e.g. Tombesi et al. 2013; Pounds & King 2013; Laha et al. 2016), and in general have small velocities or are consistent with the source rest-frame (Laha et al. 2014). The treatment of multiple detections must be care- 3 RESULTS fully considered. We wish to avoid having the sample dom- 3.1 Correlation analysis inated by a small number of sources with multiple de- tections of the same out ow at di erent velocities. For We show the UFO velocity against re ection inclination example, the UFO in PDS 456 appears at slightly dif- points in Figs 1 and 3, overplotted with simple models (see ferent velocities (from  0:23{0.33c) in almost every ob- Section 3.2). We use a Monte-Carlo approach to estimate the servation taken of the source (Go ord et al. 2014), most signi cance of any correlation between the two parameters. likely due to a ux-dependent velocity shift (Matzeu et al. The distributions of v and i are approximately log-normal 2017).However, an additional layer of ultra-fast absorption and normal, respectively, so we randomly draw 100,000 sets is present in PDS 456 (Reeves et al. 2018b) simultaneously, of points from distributions with the same mean and stan- which should be included separately. This is further compli- dard deviation (Mean log(v)= 0:89,  = 0:29, mean log(v) cated by the transient behaviour of out owing absorption i = 44:4,  = 9:53). Of these simulated sets of points, we lines (e.g. Cappi et al. 2009), which is likely due to the gas nd 37 that exceed the Pearson correlation coecient of the being fully ionized at high source uxes (Pinto et al. 2018) real data (0.64) and 629 that exceed the Spearman correla- and inhomogeneities in the wind. To mitigate this, we com- tion coecient (0.52). This gives probabilities of 0.0004 and bine multiple UFO detections from a single source into an 0.006 for a correlation this strong occurring from randomly averaged value when the velocity estimates are within the distributed points. Excluding the two sources where the re- reported errors of each other, within 0.01c of each other, or ection modeling does not meet the Reynolds (2014) crite- within 10% of each other, whichever is largest. Where the ria marginally strengthens the correlation (Pearson r =0.63, same authors have written multiple papers on a particular P = 0:0011, Spearman r = 0:49, P = 0:017). source, taking multiple velocity measurements, we assume that the later papers supersede earlier ones, and only use the latest values unless the out ows presented are clearly 3.2 Models distinct. Otherwise, we include all values in our analysis. Regardless of whether a linear correlation is a signi cant im- provement over the null hypothesis of randomly distributed points, it is still possible to use these points to infer some- 2.2 Relativistic re ection thing about the geometry of UFOs. We construct a simple Obtaining a corresponding set of re ection measurements imitation of a stream-line, where a thin out ow starts mov- is simpler than the UFO case, as there should only be one ing vertically from the disc at some radius r , following a launch value of the inclination for each AGN. We take the latest circular path with radius r . For this purpose, the units of curve available measurement for each source, unless there is an radius are arbitrary. Once it reaches a nal inclination i final earlier NuSTAR paper, in which case we use that value (in it leaves the circular path and travels on the tangent with general, NuSTAR's high energy coverage gives a better con- i = i . We also assume that the UFO accelerates along final straint on the re ection spectrum than soft instruments). this path, following v = v (1 R =(R + r )) (adapted from inf v v Reynolds (2014) gives a list of `quality control' crite- Knigge et al. 1995; Sim et al. 2008), where is a constant ria for AGN spin measurements based on re ection spec- (and = 0 gives a constant velocity) and R is a charac- troscopy. In brief, Reynolds suggests that: a full ionized re- teristic length scale. We assume that the X-ray source is ection model must be used; the iron abundance must be a coincident with the black hole (h = 0, r = 0), and ignore all free parameter; the inclination parameter must be free, and relativistic e ects. This geometry is shown in Fig. 2. This ge- constrained; and the emissivity index must be free to vary ometry is intended as an approximation of that in radiation- Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 Constraining out ow geometries 3 Table 1. Out ow velocities and inclinations for the sources in our sample. Velocities with multiple references are the result of taking the weighted average of multiple measurements that meet the criteria discussed in Section 2.1. Similarly, sources with multiple velocities are those where multiple measurements did not meet the criteria for merging. Name v (c) Reference i (degrees) Reference UFO +7:0 1H 0419-577 0:079  0:007 Tombesi et al. (2011b) 49:0 Walton et al. (2013) 4:0 z +0:01 1H 0707-495 0:11 Dauser et al. (2012) 43:0  2:0 Kara et al. (2015) 0:02 0:18  0:01 Hagino et al. (2016); Dauser et al. (2012) z y 3C 111 0:105  0:006 Go ord et al. (2013); Tombesi et al. (2011a) 44:0  2:0 Tombesi et al. (2011a) +5:0 Ark 120 0:29  0:02 Tombesi et al. (2011b) 45:0 Garc a et al. (2014) 2:0 IC 4329A 0:098  0:004 Tombesi et al. (2011b) 35:0  5:0 Mantovani et al. (2014) z +0:007 +3 IRAS 00521-7054 0:403 Walton et al. (in prep) 63 Walton et al. (in prep) 0:006 2 IRAS 13224-3809 0:236  0:006 Parker et al. (2017) 59:0  1:0 Parker et al. (2017) +2:0 IRAS 13349+2438 0:13  0:01 Parker et al. (submitted) 48:0 Parker et al. (submitted) 1:0 MCG-5-23-16 0:116  0:004 Tombesi et al. (2011b) 51:0  7:0 Zoghbi et al. (2017) +3:0 MR 2251-178 0:137  0:008 Go ord et al. (2013) 24:0 Nardini et al. (2014) 5:0 Mrk 1044 0:10  0:01 Mallick et al. (submitted) 47  3 Mallick et al. (submitted) +5:0 Mrk 509 0:14  0:0024 Cappi et al. (2009); Tombesi et al. (2011b) 50:0 Walton et al. (2013) 3:0 0:171  0:003 Cappi et al. (2009); Tombesi et al. (2011b) 0:197  0:005 Cappi et al. (2009); Tombesi et al. (2011b) +6:0 Mrk 766 0:039  0:03 Go ord et al. (2013) 39:0 Buisson et al. (submitted) 3:0 0:082  0:006 Tombesi et al. (2011b) Mrk 79 0:092  0:004 Tombesi et al. (2011b) 24:0  1:0 Gallo et al. (2011) +6:0 Mrk 841 0:055  0:025 Tombesi et al. (2011b) 46:0 Walton et al. (2013) 5:0 NGC 4051 0:202  0:006 Tombesi et al. (2011b) 37:0  5:0 Risaliti et al. (in prep) NGC 4151 0:0452  0:0099 Go ord et al. (2013); Patrick et al. (2012) < 10 Beuchert et al. (2017) 0:106  0:007 Tombesi et al. (2011b) +0:1 NGC 5506 0:246  0:006 Go ord et al. (2013) 41:0 Sun et al. (2017) 0:2 PDS 456 0:278  0:003 Reeves et al. (2018b); Matzeu et al. (2017) 65:0  2:0 Chiang et al. (2017) 0:46  0:02 Reeves et al. (2018b) PG 1211+143 0:0598  0:00069 Pounds et al. (2016); Danehkar et al. (2018); 44:0  2:0 Lobban et al. (2016) Reeves et al. (2018a) 0:129  0:002 Pounds et al. (2016) 0:151  0:003 Tombesi et al. (2011b) Swift J2127 0:231  0:006 Go ord et al. (2013) 49:0  2:0 Marinucci et al. (2014) These results do not meet the quality-control criteria of Reynolds (2014). These sources have joint re ection/UFO tting, which either gives the result reported here or is consistent with it. driven winds (e.g. Proga et al. 2000), as these give a simple locity of the absorption line and the viewing angle is to pro- explanation for the higher velocities at higher inclinations. duce the absorption in the disc itself. This model, explored MHD winds predict concave stream lines (Fukumura et al. by Gallo & Fabian (2011, 2013), explains UFO absorption 2010), so give higher velocities at small inclinations. How- using a surface layer on the disc, with the strong blueshift ever, the exact pattern observed depends on the ionization due to the orbital velocity of the absorbing material, rather and density structure within the wind, so it may still be than an out owing wind. In this case, the inclination depen- possible to explain these results in an MHD scenario. dence of the absorption velocity arises from the increased LOS velocity of the disc at high inclinations. As a simple In most cases where the line of sight intersects the wind proxy for the velocity of an absorption line from the surface in this model, it crosses the wind twice. Once while the wind of the disc, we take the maximum blueshift from the relline is rising steeply, and once in the tail where the gradient is model (Dauser et al. 2010). This gives a simple correlation constant. Because of the radial acceleration assumed and between i and v (shown in Fig. 3), although it does not reach close alignment with the line of sight in the tail, this inter- high enough velocities to account for the most blueshifted section results in a much higher apparent velocity. A simple absorption lines. The only parameter of this model is the way of only producing one measurable value for the velocity black hole spin (a), but this has a limited e ect as the max- is to assume that the gas where the rst intersection with the imum blueshift of the disc comes from further out than the LOS occurs is fully ionized. In this case, no absorption lines innermost stable circular orbit. We note that this predicted would be produced, and only the second intersection would velocity should be an upper limit as it is the maximum found be observed. An example that provides a reasonable match on the disc, so points should generally lie below the line. to the data is shown in Fig. 1, with parameters r = 10, launch r = 300, v = 0:5c, and R = 1000. We show the e ect curve v inf of varying the acceleration coecient and nal inclination 4 DISCUSSION i . From this it is clear that meaningful constraints on final these parameters can be obtained. We note that there are other possible explanations for broad An alternative way of giving a relation between the ve- emission lines in AGN. For example, Nardini et al. (2015) Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 r curve LOS final 4 M. L. Parker et al. 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 inclination (degrees) inclination (degrees) Figure 1. UFO velocity as a function of re ection inclination, overplotted with a toy model for an out owing disc wind. Left: the e ect of changing the acceleration coecient to 0.5 (red,left), 1.0 (blue, middle) and 1.5 (green, right) with nal inclination i = 65 . final Right: the e ect of setting i to 45 (red, left), 65 (blue, middle) and 85 (green, right) with = 1. final 500 interpret the emission line in PDS 456 as a P-Cygni pro- le, where the Fe K emission is produced by scattering o the out owing wind. In this case, the relativistic broad- ening is produced by the velocity of the wind instead of the orbits in the disc (Done et al. 2007). This model relies on partial-covering absorption to explain most of the spec- tral complexity. However, Chiang et al. (2017) present an alternative model where the broad-band spectrum can be fully explained by relativistic re ection, warm absorption, launch and the UFO. This interpretation is supported by the de- tection of a soft X-ray lag, generally interpreted as rever- beration in the inner disc and found in many unobscured 0 100 200 300 400 500 600 700 sources (De Marco et al. 2013). While the detection of an X-ray lag is usually considered strong evidence for the pres- ence of relativistic re ection in a source (and many of the Figure 2. Simple streamline geometry, with the wind shown in sources in our sample show reverberation lags, Kara et al. blue and the line of sight in red (dashed). Length units are arbi- 2016; De Marco et al. 2013), it is dicult to rule out a con- trary. tribution to the total broad-line pro le from scattering in the wind. Indeed, given that disc winds are likely to be most 0.5 dense at the point they launch from the disc and should be a = 0.998 co-rotating with it, then scattering from the disc surface and a = 0 wind may be thought of as a single continuous process. The 0.4 impact of having Fe K emission from both the disc and wind is not understood, but could have a signi cant e ect on the measured inclination. A higher velocity out ow will produce 0.3 more blueshifted emission, in the same way that a higher in- clination gives a more blueshifted line pro le for disc re ec- tion. We note that the tentative relation identi ed here may 0.2 be indicative of the e ect of wind emission on the net rel- ativistic line pro le rather than an inclination dependence. 0.1 A related issue is the lack of simultaneous UFO/re ection modeling. Using a simpli ed phenomenological continuum may exaggerate the signi cance of line features, leading to 0.0 false detections (Zoghbi et al. 2015). Similarly, not account- 0 10 20 30 40 50 60 70 80 inclination (degrees) ing for UFO absorption during re ection modeling may bias the measured parameters. We will revisit the spectra of some Figure 3. Toy model for absorption from a layer on the disc. of these sources for joint modeling to investigate this further in future work. We have assumed throughout that all UFOs have the Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018 v/c v/c v/c Constraining out ow geometries 5 same shape and velocity pro le, and that the observed ve- Fukumura K., Tombesi F., Kazanas D., Shrader C., Behar E., Contopoulos I., 2015, ApJ, 805, 17 locity is solely determined by the LOS angle. This is unlikely Gallo L. C., Fabian A. C., 2011, MNRAS, 418, L59 to be the case in practice: the accretion rate, for example, is Gallo L. C., Fabian A. C., 2013, MNRAS, 434, L66 likely to have a major impact on the velocity of the material. Gallo L. C., Miniutti G., Miller J. M., Brenneman L. W., Fabian Similarly, we have implicitly assumed that wind instabilities A. C., Guainazzi M., Reynolds C. S., 2011, MNRAS, 411, 607 play a negligible role in determining the observed velocity, Garc a J., et al., 2014, ApJ, 782, 76 which is unlikely. Another caveat is that our sample is any- George I. M., Fabian A. C., 1991, MNRAS, 249, 352 thing but unbiased, and the biases are poorly understood. Go ord J., Reeves J. N., Tombesi F., Braito V., Turner T. J., Re ection measurements are generally biased towards high Miller L., Cappi M., 2013, MNRAS, 430, 60 spin (Vasudevan et al. 2016), and there may be a similar Go ord J., et al., 2014, ApJ, 784, 77 bias towards high inclination, as it produces broader, easier Hagino K., Odaka H., Done C., Tomaru R., Watanabe S., Taka- hashi T., 2016, MNRAS, 461, 3954 to measure lines. Similarly, it is plausible that UFO veloc- Kara E., et al., 2015, MNRAS, 449, 234 ity measurements are biased towards lower velocities, as the Kara E., Alston W. N., Fabian A. C., Cackett E. M., Uttley P., sensitivity of X-ray detectors typically declines with energy Reynolds C. S., Zoghbi A., 2016, MNRAS, 462, 511 (although this may be remedied as the NuSTAR archive Knigge C., Woods J. A., Drew J. E., 1995, MNRAS, 273, 225 grows). Laha S., Guainazzi M., Dewangan G. C., Chakravorty S., Kemb- havi A. K., 2014, MNRAS, 441, 2613 Laha S., Guainazzi M., Chakravorty S., Dewangan G. C., Kemb- 5 CONCLUSIONS havi A. K., 2016, MNRAS, 457, 3896 Lobban A. P., Pounds K., Vaughan S., Reeves J. N., 2016, ApJ, We have identi ed a correlation between the velocity of 831, 201 highly ionized absorption features from UFOs and the in- Madej O. K., Garc a J., Jonker P. G., Parker M. L., Ross R., clination of the inner accretion disc measured from re ec- Fabian A. C., Chenevez J., 2014, MNRAS, 442, 1157 tion spectroscopy. The correlation is formally signi cant, but Mantovani G., Nandra K., Ponti G., 2014, MNRAS, 442, L95 Marinucci A., et al., 2014, MNRAS, 440, 2347 heavily reliant on a small number of points at high velocity. Matzeu G. A., Reeves J. N., Braito V., Nardini E., McLaughlin We show that the observed points can be explained by D. E., Lobban A. P., Tombesi F., Costa M. T., 2017, MNRAS, simple toy models of an out owing wind or absorption from 472, L15 a disc, although the latter cannot account for the highest Middleton M. J., Parker M. L., Reynolds C. S., Fabian A. C., velocity features. With more detailed modeling and higher Loh nk A. M., 2016, MNRAS, 457, 1568 quality data, this technique could be very powerful for con- Nardini E., Reeves J. N., Porquet D., Braito V., Grosso N., Gof- straining the geometry of out owing material in AGN. ford J., 2014, MNRAS, 440, 1200 Nardini E., et al., 2015, Science, 347, 860 Parker M. L., et al., 2017, Nature, 543, 83 Patrick A. R., Reeves J. N., Porquet D., Markowitz A. G., Braito ACKNOWLEDGEMENTS V., Lobban A. P., 2012, MNRAS, 426, 2522 MLP and GM are supported by European Space Agency Pinto C., et al., 2018, MNRAS, 476, 1021 (ESA) Research Fellowships. JJ acknowledges support by Pounds K. A., King A. R., 2013, MNRAS, 433, 1369 the Cambridge Trust and the Chinese Scholarship Coun- Pounds K. A., Reeves J. N., Page K. L., Wynn G. A., O'Brien P. T., 2003, MNRAS, 342, 1147 cil Joint Scholarship Programme (201604100032). DJKB is Pounds K., Lobban A., Reeves J., Vaughan S., 2016, MNRAS, supported by an STFC studentship. 457, 2951 Proga D., Stone J. M., Kallman T. R., 2000, ApJ, 543, 686 Reeves J. N., O'Brien P. T., Ward M. J., 2003, ApJ, 593, L65 REFERENCES Reeves J. N., Lobban A., Pounds K. A., 2018a, ApJ, 854, 28 Reeves J. N., Braito V., Nardini E., Lobban A. P., Matzeu G. A., Beuchert T., et al., 2017, A&A, 603, A50 Costa M. T., 2018b, ApJ, 854, L8 Cappi M., et al., 2009, A&A, 504, 401 Reynolds C. S., 2014, Space Sci. Rev., 183, 277 Chartas G., Brandt W. N., Gallagher S. C., Garmire G. P., 2002, Sim S. A., Long K. S., Miller L., Turner T. J., 2008, MNRAS, ApJ, 579, 169 388, 611 Chiang C.-Y., Cackett E. M., Zoghbi A., Fabian A. C., Kara E., Sun S., Guainazzi M., Ni Q., Wang J., Qian C., Shi F., Wang Y., Parker M. L., Reynolds C. S., Walton D. J., 2017, MNRAS, Bambi C., 2017, preprint, (arXiv:1704.03716) 472, 1473 Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Yaqoob Danehkar A., et al., 2018, ApJ, 853, 165 T., Braito V., Dadina M., 2010, A&A, 521, A57 Dauser T., Wilms J., Reynolds C. S., Brenneman L. W., 2010, Tombesi F., Sambruna R. M., Reeves J. N., Reynolds C. S., Braito MNRAS, 409, 1534 V., 2011a, MNRAS, 418, L89 Dauser T., et al., 2012, MNRAS, 422, 1914 Tombesi F., Cappi M., Reeves J. N., Palumbo G. G. C., Braito De Marco B., Ponti G., Cappi M., Dadina M., Uttley P., Cackett V., Dadina M., 2011b, ApJ, 742, 44 E. M., Fabian A. C., Miniutti G., 2013, MNRAS, 431, 2441 Tombesi F., Cappi M., Reeves J. N., Nemmen R. S., Braito V., Done C., Sobolewska M. A., Gierlinsk  i M., Schurch N. J., 2007, Gaspari M., Reynolds C. S., 2013, MNRAS, 430, 1102 MNRAS, 374, L15 Vasudevan R. V., Fabian A. C., Reynolds C. S., Aird J., Dauser Fabian A. C., 2012, ARA&A, 50, 455 T., Gallo L. C., 2016, MNRAS, 458, 2012 Fabian A. C., Rees M. J., Stella L., White N. E., 1989, MNRAS, Walton D. J., Nardini E., Fabian A. C., Gallo L. C., Reis R. C., 238, 729 2013, MNRAS, 428, 2901 Fabian A. C., et al., 2009, Nature, 459, 540 Zoghbi A., et al., 2015, ApJ, 799, L24 Fukumura K., Kazanas D., Contopoulos I., Behar E., 2010, ApJ, Zoghbi A., et al., 2017, ApJ, 836, 2 715, 636 Downloaded from https://academic.oup.com/mnrasl/advance-article-abstract/doi/10.1093/mnrasl/sly096/5026625 by Ed 'DeepDyve' Gillespie user on 08 June 2018

Journal

Monthly Notices of the Royal Astronomical Society LettersOxford University Press

Published: Jun 1, 2018

Keywords: accretion, accretion discs; black hole physics; galaxies: active

There are no references for this article.