Skip to main content

Full text of "Broad Iron Emission from Gravitationally Lensed Quasars Observed by Chandra"

See other formats


arXiv:1503.05255vl [astro-ph.HE] 18 Mar 2015 


Draft version March 19, 2015 

Preprint typeset using style emulateapj v. 5/2/11 


BROAD IRON EMISSION FROM GRAVITATIONALLY LENSED QUASARS OBSERVED BY CHANDRA 

D. J. Walton*'^ M. T. REYNOLDS^ J. M. MILLER^ R. C. REIS^ D. Stern*, F. A. Harrison^ 

* Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109, USA 
^ Space Radiation Laboratory, California Institute of Technology, Pasadena, CA 91125, USA 
^ Department of Astronomy, University of Michigan, 1085 S. University Ave., Ann Arbor, MI, 49109-1107, USA 

Draft version March 19, 2015 

ABSTRACT 

Recent work has demonstrated the potential of gravitationally lensed quasars to extend measurements of 
black hole spin out to high-redshift with the current generation of X-ray observatories. Here we present an 
analysis of a large sample of 27 lensed quasars in the redshift range 1.0 z < 4.5 observed with Chandra, 
utilizing over 1.6 Ms of total observing time, focusing on the rest-frame iron K emission from these sources. 
Although the X-ray signal-to-noise (S/N) currently available does not permit the detection of iron emission 
from the inner accretion disk in individual cases in our sample, we find significant structure in the stacked 
residuals. In addition to the narrow core, seen almost ubiquitously in local AGN, we find evidence for an 
additional underlying broad component from the inner accretion disk, with a clear red wing to the emission 
profile. Based on simulations, we find the detection of this broader component to be significant at greater than 
the 3 ct level. This implies that iron emission from the inner disk is relatively common in the population of 
lensed quasars, and in turn further demonstrates that, with additional observations, this population represents 
an opportunity to significantly extend the sample of AGN spin measurements out to high-redshift. 

Subject headings: Black hole physics - Galaxies; active 


1. INTRODUCTION 

Information regarding the manner in which the supermas- 
sive black holes (SMBHs) powering active galactic nuclei 
(AGN) grew is encoded in the distribution of their angu¬ 
lar moment a J (or more specifically their ‘spin’, a* _ = 

Jc/GMpTT-. iKing & Prindel 120061 iBerti & Volonteril 120081 
IVolonteri et al.ll2013l) . For Sample, if these black holes grew 
through prolonged episodes of coherent accretion, we should 
see a preference for rapidly rotating black holes. In contrast, 
if they grew through a series of chaotic mergers and accretion 
events we should instead see a lower value for the average 
black hole spin. 

AGN spin measurements are anchored in X-ray spec¬ 
troscopy, and rely primarily on measuring the relativistic dis- 
tortio ns imprinted on fluorescent lin e emission from the inner 
disk (iFabian et al.lll98^ lLaodll99l1) . which result in intrinsi¬ 
cally narrow emission lines being broadened and skewed into 
a characteristic ‘diskline’ profile. Although lines from a va¬ 
riety of elements are naturally produced in the resulting ‘re¬ 
flected’ emission when the optically thick accretion disk is 
irradiated by high-energy X-rays, the most prominent is the 
iron Ka line at ~6-7 keV (depending on ionisation state), 
owing to its high cosm ic abundance and fluorescent yield 
(iGeorge & Fabianlll991h . It is also best suited for measur¬ 
ing these relativistic distortions, being relatively isolated from 
other strong emissi on lines. For rece nt reviews on rela tivistic 
disk reflection, see iMillerl (120071) and iRevnoldsI (l20 1 3l) . 

Spin estimates for a growing sample of ^20-30 local (z ^ 
0) A GN have been obtained through study of t hese features, 
(e.g. IWalton et all 120131 IFabian et al.l 120131: iRisaliti et al.l 
120131) . and already suggest that many black holes have high 
spin, althoug h the sample is not yet well define d in a sta¬ 
tistical sense (iBrennemanI |20 1 31 iRevnold^ 120131) . As mod¬ 
els linking the cosmic growt h of SMBHs and galaxy forma- 
tion become more advanced (iDubois, Volonteri & Silkll2014l 


ISesana et alJ2014l) . detailed comparison with observation will 
require knowledge of the spin distribution as a function of red- 
shift, e xtending out to and b eyond the peak of AGN activity 
(z ^ 2: lRichards et al.ll2006h . However, current X-ray instru¬ 
mentation does not have the sensitivity required to undertake 
such measurements for typical AGN at these redshifts. 

Recently, however, we have demonstrated that strongly 
lensed quasars offer a rare opportunity to obtain spin mea¬ 
surements from objects at cosmologically interesting redshifts 
with current instrumentation, owing to the combination of the 
multiple images observed and the amplification of the intrin¬ 
sic emission by the lens. The two cases with sufficient signal- 
to-noise (S/N), RXJl 131-1231 (z = 0.658) and Q2237H-0305 
(aka the Einstein Cross; z = 1.695), exhibit relativistic iron 
disklines similar to those observed in local Seyfert galaxies, 
allowing us to infer that both sources host rapid ly rotating 
black holes (iReis et alJl2014l: [Reynolds et alJl201^ . Here, we 
examine the iron emission from a sample of lensed quasars 
with X-ray observations that currently have lower S/N, to in¬ 
vestigate whether such iron disklines are common among this 
population, and whether it could potentially facilitate further 
high-redshift spin measurements. 

2. SAMPLE SELECTION AND DATA REDUCTION 

We selected a sample of strong gravitationally lensed 
quasars from the publicly available CASTLES databasf] 
(numbering 100 lensed quasars in total as of August 2014). 
We select all those systems characterized as grade A (i.e., 
confirmed multiply imaged strong gravitationally lensed 
quasar^ with redshift determinations for both the lens and 
the quasar (e.g., there are eight grade A systems with no se¬ 
cure redshift for the lens, which we exclu de from our anal- 
ysis). Additionally, both RXJ1131-1231 (iReis et akllMl) 

^ http://WWW.cfa.harvard.edu/castles/ 

^ per the CASTLES website: “Fd bet my life on it”. 












































2 


D. J. Walton et al. 


and Q2237+0305 (iRevnolds et al.llMll are excluded from 
the sample, having individually been shown to exhibit promi¬ 
nent relativistic lines (see also iDai et al] 120031: IChen et aH 
l2012h . We then searched for sources with i maging observa¬ 
tions obtained with the Chandra observatory (iWeisskopf et al.l 
l2002h . restricting ourselves to obs ervations within 1' o f the S3 
aimpoint on the ACIS detector (iGarmire et al.ll200^ . This 
resulted in an initial sample of 35 systems. The major¬ 
ity of this sample are lensed by foreground galaxies, but 
three sources are lensed by galaxy clusters (O0957-H561, 
SDSSJ1004-H4112 an d S DSSJ1029+2623 ; see iChartas et alJ 
IMillOtaet dllMia and lOta et al.ll20l:^ respectively), and 
two are lensed by galaxies within faint X-ray emitting g alaxy 
groups (PGl 115H-080 and B1422H-231 : lGrant et alJ200l . We 
note however that previous work has shown that the clus¬ 
ter/group emission peaks at <lkeV in the observed frame, 
and that it makes a negligible contribution to the iron band¬ 
pass for the quasars in these cases. 

All observations were re-processed in CIAO v4. 5a with 
the EDSER algorithm enabled and with the latest Chandra 
calibration files. The resulting event files were re-binned to 
1/8*** of the native ACIS pixel size before smoothing with a 
Gaussian of 0.25" (EWHM). In those cases where individ¬ 
ual images were resolved, spectra were extracted from 0.5" 
radius regions centered on each individual sub-pixel image 
via the SPECEXTRACT script with PSE correction enabled. If 
this was not possible (e.g. due to the quasar images being too 
close for even Chandra to resolve, or low count rates preclud¬ 
ing centroiding), larger extraction regions encompassing mul¬ 
tiple images were utilized. All spectra for each source were 
subsequently combined using the COMBINE_SPECTRA script 
before grouping to S/N=3 per spectral bin with DMGROUP; 
we verified that we obtain consistent results using an initial 
binning of S/N=4. The spectra and background files for all 
sources we re then export ed to XSPEcQ for spectral analysis 
(V12.8.Om: I Arnaudll 1 9961) . In the observed frame, the Chan¬ 
dra data are modeled over the 0.35-8.0keV bandpass. 

3. STACKING ANALYSIS 

To investigate the iron K emission from our sample we 
stacked the 2-lOkeV data to produce a single, average spec¬ 
trum, owing to the low photon statistics currently available 
for the individual sources in our sample (see Table [T]i, adopt¬ 
ing an approach similar to the r esidua ls-based methodol¬ 
ogy outli ned in iChaud hary et al.l (120121 ), as described be¬ 
low ( cf. iNandra et al.l 119971 iGuainazzi, Bianchi & DovciakI 
|200^. We limit ourselves to this method, rather than addi¬ 
tionally redshift-correcting and stacking the spectra an d in¬ 
strumental responses directly, as IChaudharv et al.l (12012h find 
in their analysis of 2XMM sources that the two methods re¬ 
turn consistent results. Eurthermore, this method has the ad¬ 
vantage that the extinction towards each source can be treated 
individually. 

Eor each individual source we fit a simple absorbed pow- 
erlaw continuum to the full observed bandpass, excluding the 
rest-frame 3.5-7.5keV energy range where the iron emission 
may contribute, and then determined the data/model ratio for 
the 2-lOkeV rest-frame bandpass. This model includes both 
a fixed Galactic absorption component (iKalberla et al.ll2005h 
and a second neutral absorption component at the redshift of 
the source that was free to vary, and we limited the photon 

^ http://cxc.harvard.edu/ciao 

^ http://heasarc.gsfc.nasa.gov/xanadu/xspec/ 


TABLE 1 

The sample oe 27 gravitationally lensed quasars observed by 
Chandra included in our analysis. 


Source 

^qso 

"^lens 

TV;” 

Total 

exposure^^ 

2-10 keV 
counts^ 

B 1152-H199 

1.019 

0.439 

2 

27 

1927 

SDSS J1226-0006 

1.12 

0.52 

2 

5 

48 

EBQ 0951-1-2535 

1.24 

0.26 

2 

35 

132 

Q 0158-4325 

1.29 

0.317 

2 

35 

550 

B 0712-H472 

1.34 

0.41 

4 

99 

627 

SBS 0909-H532 

1.377 

0.83 

2 

20 

1520 

Q 0957-H56H 

1.41 

0.36 

2 

37 

10919 

SDSS J0924-I-0219 

1.524 

0.39 

2 

128 

735 

B 1600-H434 

1.59 

0.41 

2 

31 

170 

HE 0047-1756 

1.66 

0.41 

2 

40 

1017 

WFI12033-4723 

1.66 

0.661 

4 

30 

378 

HE 0435-1223 

1.689 

0.46 

4 

no 

1699 

PG 1115-1-080 

1.72 

0.31 

4 

156 

6353 

SDSS 11004-1-4112“' 

1.734 

0.68 

4/5“ 

155 

7745 

HE 0230-2130 

2.162 

0.52 

4 

45 

1063 

SDSS 11029-h2623‘' 

2.197 

0.55 

2 

57 

1715 

PMN 10134-0931 

2.216 

0.77 

4/5“ 

1 

73 

HE 1104-1805 

2.32 

0.73 

2 

131 

2192 

SDSS11138-H0314 

2.438 

0.445 

4 

51 

160 

MG 10414-1-0534 

2.639 

0.9854 

4 

200 

8696 

QSO 11004-1-1229 

2.65 

0.95 

2 

19 

36 

Q 0142-100 

2.72 

0.49 

2 

14 

407 

LBQS 1009-0252 

2.74 

0.87 

2 

10 

94 

RX 10911.4-1-0551 

2.80 

0.77 

4 

65 

347 

HS 0818-1-1227 

3.115 

0.39 

2 

20 

197 

B 1422-1-231 

3.62 

0.34 

4 

125 

9509 

BRI 0952-0115 

4.5 

0.632 

2 

20 

42 

Total 




1666 

58351 

“ The number of lensed images of each 

quasar included in our sample 


^ Total Chandra exposure in ks 
Assessed in the quasar rest-frame 
Lensed by a cluster of galaxies 

Although there are five lensed images in the optical, only four have 
robust X-ray detections. 

index to 1.3 < E < 3.0. Of the 35 sources in our initial sam¬ 
ple, 8 did not have sufficient data to fit the continuum above 
and below the iron bandpass. These were subsequently ex¬ 
cluded from our analysis, resulting in a final sample of 27 
lensed QSOs; the basic details of this sample are presented in 
Table □ 

We then corrected the ratio spectra obtained for the source 
redshift, and re-sampled to a common rest-frame energy bin¬ 
ning. When re-sampling the ratio spectra, we assumed the 
counts included in each of the original bins to be evenly 
distributed across the energy range covered by that bin, and 
present the results with this analysis. However, identical re¬ 
sults are obtained if we instead assume a powerlaw distri¬ 
bution, adopting the photon index obtained from our initial 
continuum fits for each source respectively. To determine 
the data/model ratio for each of the new bins, the ratios in 
each of the overlapping original bins were weighted by the 
fraction of the energy range of the new bin that they con¬ 
tribute, and then averaged. The fractional uncertainties were 
determined from the counts in each new bin, assuming Pois¬ 
son statistics. We re-sampled the 2-lOkeV bandpass into 50 
evenly spaced energy bins of width 160eV, slightly broader 















































Broad Iron Emission from Gravitationally Lensed Quasars 


3 



Energy (keV) 


Fig. 1.— The stacked data/model residuals in the Fe K bandpass for the 
sample of 27 lensed quasars observed by Chandra, modeled outside the Fe 
K bandpass with a simple absorbed powerlaw continuum. The stacked iron 
emission profile exhibits a combination of a narrow core on top of a broader 
diskline-like component. 


than the ACIS spectral resolution at the observed iron ener¬ 
gies, given the redshift range probed. Einally, we stacked the 
re-sampled ratio spectra by calculating the weighted average 
of the individual ratios in each of the new bins. The total 
number of counts contributing to each of our re-sampled en¬ 
ergy bins is always greater than 200. In order to test our re¬ 
sampling/stacking analysis, we simulated a series of spectra 
at different redshifts with a common model, consisting of a 
powerlaw continuum and a diskline emission line profile, ap¬ 
plied our analysis to these simulated spectra, and verified that 
the stacked residuals reproduced the input line profile. We 
also confirmed that the results presented below do not depend 
on our re-sampling, again obtaining consistent results with an 
analysis re-sampling to 40 bins (200 eV width) instead. 

4. IRON EMISSION 

The profile of the stacked iron residuals obtained with our 
analysis is shown in Eigure [T] It reveals a combination 
of a narrow core, seen almost ubiquitously in local AGN 
(iNandra et al.ll200% . on top of a broader diskline-like compo¬ 
nent. In order to analyse the stack ed iron emission profile, we 
followed IChaudharv et al.l (1201 2h and generated a powerlaw 
continuum with the same energy binning as used above, then 
multiplied this continuum by the observed ratio spectrum. We 
adopted a photon index of E = 1.8 for the continuum, the me¬ 
dian of the distribution obtained from our powerlaw fits to the 
observed sample. This was then imported into XSPEC using 
FLX2XSP (part of the HEASOFT distribution). 

A simple powerlaw continuum provides a poor fit to the re¬ 
sulting spectrum (x^/DoF = 109/48). We next include a 
narrow emission line from neutral iron. However, given the 
range of redshifts, the rest-frame iron Ka lines occur at a 
range of observed energies. Differences in the spectral res¬ 
olution of the ACIS detectors at these energies will result in 
some apparent broadening of narrow features beyond the in¬ 
strumental resol ution at 6 keV when s tacking spectra from dif¬ 
ferent redshifts (llwasawa et al.ll20T^ . Furthermore, our bin¬ 
ning of the individual spectra, along with our assumption that 


T 



Energy (keV) 


Fig. 2.— The model applied to the stacked 2-iOkeV residuals after the 
re-introduction of a F = 1.8 powerlaw continuum (top) and the data/model 
residuals with this model applied (middle). The total model is shown in black, 
and the powerlaw continuum in red; the inset shows a zoom-in on the model 
for the Fe K profile. The parameters adopted for the broad line component 
are a* = 0.7, i = 45°, qin = 6, pout = 3, i?br = 6 Rq (see text), and we 
find a best-fit energy of E 6.7 keV. The data/model residuals for a model 
including just a narrow iron emission line are also shown (bottom). 


photons within these bins are evenly distributed as a function 
of energy may also result in some broadening of narrow fea¬ 
tures. In order to assess the combination of these effects we 
performed a series of simulations (see section|5]l, and find that 
our analysis procedure should cause an intrinsically narrow 
iron emission line to have a width of cr = 0.165keV in the 
stacked spectrum. We therefore fix the width of the narrow 
line to 0.165keV in our analysis, and we also fix the energy 
to that of neutral iron (6.4keV). The addition of this line im¬ 
proves the fit (x^/DoF = 85/47) and successfully accounts 
for the narrow core of the observed emission profile. How¬ 
ever, broad residuals still remain (see Figure|2l). 

We therefore include a second, broad Gaussian line, which 
further improves the fit (x^/DoF = 61/44). The broad com¬ 
ponent has E = 6.2 ± O.SkeV and a = l.O/^Q gkeV (errors 
are 90% confidence), significantly larger than the combined 
broadening of the instrument and our analysis. We also con¬ 
struct a model in which the broad Gaussian line is replaced by 
a relativistic emission line from the innermost accretion disk, 
using the RELLINE model (iDauser et al.ll20ldh . Although we 
are adopting a simplistic approach, modeling what would in 
reality be a combination of direct continuum and a full re¬ 
flection spectrum from the inner disk with a powerlaw con¬ 
tinuum and a single relativistic emission line, given the S/N 
more detailed treatments are not warranted, but it is worth 
noting that the former will also incorporate some contribution 
from the blurred reflection continuum that would naturally 
accompany the line. For consistency, the RELLINE compo- 







































































































4 


D. J. Walton et al. 


nent is smoothed with a a = 0.165keV Gaussian, as per the 
broadening of narrow features in our stacked spectrum. In this 
model, the details of the line profile are determined by several 
parameters, and given the signal we are not able to indepen¬ 
dently constrain them all. We therefore fix several of these 
parameters to physically motivated values in our analysis, but 
we stress that the results obtained do not strongly depend on 
our adopted parameter values. 

The parameters we fix are the black hole spin, and the incli¬ 
nation and emissivity profile of the accretion disk. We set the 
spin to be a* = 0.7, motivate d by both our previous results for 
high- redshift lensed quasars (iReis et alJl2014t [Reynolds et alJ 
|20T3), and the spin impl ied from the re lation between BH spin 
and radiative efficiency (iThornd 1974ft an d the av e rage q uasar 
radiative efficiency of p ~ 0.1 found by ISoltaiil (Il982ft . We 
further assume that the disk extends into the innermost stable 
circular orbit. The inclination is set to 45°, roughly similar 
to that expected for luminous, unobscured quasars. Finally, 
we adopt a broken powerlaw profile for the radial emissivity 
of the disk, with an inner index of qtn = 6, an outer index 
of gout = 3, and a break radius of i?br = 6i?G- This is 
motivated by both t he compact X-ray source sizes inferred 
for lensed quasars (iDai et al.l 120101 iMosauera et alJ 120131 


IReis & Milleill20l'^lMacLeod et al. 12015ft a nd the relativistic 

ray-tracing work of l Wilkins & FabianI (12012ft . The line energy 
and normalisation are free to vary, although the energy is con¬ 
strained to the range of iron Ka transitions (6.4-6.97keV). 
The inclusion of this RELLINE component formally provides 
the best fit of all the models (x^/DoF = 60/45), and the 
line energy obtained is E > 6.6 keV. The equivalent widths 
of the relativistic line and the narrow core in our redshift- 
corrected stacked spectrum are F^Wrei = 190 ± 70 eV and 
F^VFnarrow = 28^27 eV. The strength of the relativistic emis¬ 
sion is very similar t o that expected for reflec tion from a stan¬ 
dard accretion disk (iGeorge & Fabianlll991ft . In addition to 
testing different spectral binning and re-sampling methods, as 
discussed above, we have also systematically excluded each 
source and repeated the analysis on the reduced sample, and 
find that these results are robust to the exclusion of any indi¬ 
vidual source. 


5. SIMULATIONS 

In order to interpret these results, we also performed a se¬ 
ries of simulations to assess the detection significance of the 
broad component of the iron emission. Using the Chandra 
responses, we simulated the spectra for 10,000 samples of 
27 sources using EAKEIT in XSPEC, and performed the same 
stacking analysis outlined above. We simulated an absorbed 
powerlaw continuum with a narrow (cr = 10 eV) iron line 
for each source, using the continuum parameters obtained in 
our initial analysis of the real sample, an equivalent width of 
^80 eV and the accompanying Chandra exposure. No broad 
iron emission component was included. We included count¬ 
ing statistics in our simulations, and rebinned the individual 
simulated spectra to the same level as the real data before 
performing the stacking analysis described above. For each 
of the 10,000 stacked ratio spectra obtained, we analyzed the 
simulated data in the same manner as above, applying them 
to a F = 1.8 powerlaw continuum and modeling the result 
in XSPEC. We first model the stacked spectra with a pow¬ 
erlaw continuum and a Gaussian emission line at 6.4 keV. 
The line width is free to vary, in order to determine the com¬ 
bined broadening introduced by our analysis procedure and 
the range of instrumental resolutions at the observed Fe K en¬ 


ergies (owing to the different source redshifts). As discussed 
previously, we find the average width for the narrow line in the 
stacked spectra to be cr = 0.165 keV. We then add a RELLINE 
component with the parameters used above, smoothed by a 
Gaussian with the width linked to that found for the narrow 
line, and note the improvement in x^, in order to assess the 
chance probability of obtaining the observed improvement in 
the instance that no broad line is actually present. Of the 
10,000 samples simulated, only 7 show a chance improve¬ 
ment equivalent to or greater than that observed, implying that 
the detection significance of the broad feature detected in the 
real data exceeds the 3cr level. 

6. DISCUSSION AND CONCLUSIONS 

Following our recent detection of relativistic disk reflec¬ 
tion features in two lensed quasars beyond the local uni¬ 
verse, RXJl 131-1231 (z ^ 0.658: IReis et al.l IMTl) and 
Q2237H-0305 (z = 1.695: lRevnolds et al.ll2014ft 7we have pre¬ 
sented an analysis of a large sample of 27 lensed quasars in 
the redshift range 1.0 ^ z < 4.5 observed with C/ranc/ra, with 
a combined exposure of over 1.6 Ms. These do not currently 
have the archival S/N of the two individual cases presented to 
date. While iron emission has been det ected from a few o f the 
sources in our sample individually (e.g. lPage et al.ll2()()^ . and 
there have previously been low-significance claims of com¬ 
plexity in the iron bandpass f or SDSS J0924H-0219 and SDSS 
J1004H-4112 (lota et al.l200'^IChen et al.llMTl) . none have ro¬ 
bust relativistic line detections, so we focus on stacking the 
residuals in the iron K bandpass to simple po werlaw AGN 
continuum models, similar to previous analyses (iNandra et al. 
j_997 : iGuainazzi, Bianchi & Dovciakll200CT IChaudharv et al. 
2012ft . We find that the stacked iron emission from this 
sample shows both a narrow core and an underlying broad 
diskline-like component, with a clear red wing to the emis¬ 
sion profile (see Figure[Tli. 

The presence of this broad component is robust to the vari¬ 
ous aspects of our analysis, and is not dominated by the con¬ 
tribution from any one source. This is an important point, as 
some of the sample may also individually exhibit additional 
spectral complexities. For example, some Chandra observa¬ 
tions of PGl 115-1-080 may show evidence for an X-ray out¬ 
flow via hig h-energy iron absorption lines at ^7 .4 and 8.5 keV 
(rest-frame: IChartas, Brandt & Gallagherl2003ft . Fiye sources 
are lensed by either galaxy clusters or groups rather than 
indiyidual galaxies (Chartas et al.l l2002t [Grant et’aD 120041: 
lOta et al.] 120061 12012 ). although in all these cases the clus¬ 
ter/group emission is known to be negligible in the quasars 
iron bandpass. Repeating the analysis excluding these five 
sources, we find including RELLINE still gives a similar sta¬ 
tistical improvement, but the data at the highest energies has 
much lower S/N. The robustness of the detection of the broad 
component to the exclusion of any individual source in turn 
implies it is also robust to such details, and is genuinely rep¬ 
resentative of the average properties of our sample. 

We interpret this broad emission component as iron emis¬ 
sion originating from the inner accretion disks of these high- 
redshift sources. Although the S/N is still low and other inter¬ 
pretations can almost certainly reproduce the data, for exam¬ 
ple the right combination of comple x absorption components 
could mimic a diskline-like profile (iMiller, Turner & ReevesI 
12009ft . our interpretation is motivated by several recent results 
regardi ng local AGN. With the launch of the NuSTAR obser¬ 
vatory (iHarrison et al.l20r3l) . high S/N broadband X-ray stud¬ 
ies of local AGN are now possible, and have demonstrated 























































Broad Iron Emission from Gravitationally Lensed Quasars 


5 


the presence of reflection fr om the inner accretion disk in the 
Seyfe rt galaxy NGC1365 (iRisaliti et alJl201^ IWalton et alJ 
|2013)- For the highest S/N local AGN, reverb eration of the 
broad iron emission has al s o now been detec ted (IZoghbi et al.l 
120121 l^ckett et aTll2014t iKara et all 120151) . unambiguously 
demonstrating this emission arises through reprocessing of 
the continuum close to the black hole. In addition, fea¬ 
tures consistent with relativistic line emission are observed 
in both local active galaxies and Galactic black hole binaries 
(IWalton et al.ll20T2h . as expected if they arise from the inner 
disk. Finally, one of the key advantages of lensed quasars is 
that the size of the X-ray emitting region can be determined 
independently through microlensing stu dies. These typically 
find the X-ray source to be compaQ Joi artas et a l.l 1200^ 
iDai et akl 120101: iMorgan et alJ 1201^ iMosauera et al.l 120131) . 
broadly consistent with the sizes inferred from disk reflec¬ 
tion/reverberation from local sources. We stress that while 
microlensing could potentially influence the precise details 
of the iron emission profile, in particular the radial emissiv- 
ity, it cannot artificially broaden the line, as it only serves to 
magnify the intrinsic emission. Regardless, by averaging both 
over time and over the sample, such effects should naturally 
be reduced. 

Other stacking studies of moderate/high-redshift AGN in 
the i ron bandpass have also shown hints of broad emis- 
sion (iBrusa, Gilli & Comastril 120051: IChaudharv et al.1120121: 


Amaud K. A., 1996, in Astronomical Society of the Pacific Conference 
Series, Vol. 101, Astronomical Data Analysis Software and Systems V, 

G. H. Jacoby & J. Barnes, ed., p. 17 
Berti E., Volonteri M., 2008, ApJ, 684, 822 

Brenneman L., 2013, Measuring the Angular Momentum of Supermassive 
Black Holes, SpringerBriefs in Astronomy, New York: Springer 
Bmsa M., Gilli R., Comastri A., 2005, ApJ, 621, L5 
Cackett E. M., Zoghbi A., Reynolds C., Fabian A. C., Kara E., Uttley P., 
Wilkins D. R., 2014, MNRAS, 438, 2980 
Chartas G., Brandt W. N., Gallagher S. C., 2003, ApJ, 595, 85 
Chartas G., Gupta V., Garmire G., Jones C., Ealco E. E., Shapiro 1.1., 
Tavecchio R, 2002, ApJ, 565, 96 

Chartas G., Kochanek C. S., Dai X., Poindexter S., Garmire G., 2009, ApJ, 
693,174 

Chaudhary R, Bmsa M., Hasinger G., Merloni A., Comastri A., Nandra K., 
2012, A&A, 537, A6 

Chen B., Dai X., Kochanek C. S., Chartas G., Blackbume J. A., Morgan 
C. W., 2012, ApJ, 755,24 

Dai X., ChaiJas G., Agol E., Bautz M. W., Garmire G. R, 2003, ApJ, 589, 
100 

Dai X., Kochanek C. S., Chartas G., Kozlowski S., Morgan C. W., Garmire 
G., Agol E., 2010, ApJ, 709, 278 

Dauser T., Wilms J.. Reynolds C. S., Brenneman L. W., 2010, MNRAS, 409, 
1534 

Dubois Y, Volonteri M., Silk J., 2014, MNRAS, 440, 1590 
Fabian A. C. et al., 2013, MNRAS, 429, 2917 

Fabian A. C., Rees M. J., Stella L., White N. E., 1989, MNRAS, 238, 729 
Falocco S. et al., 2013, A&A, 555, A79 

Garmire G. P. Bautz M. W., Ford P. G., Nousek J. A., Ricker, Jr. G. R., 

2003, in SPIE Conference Series, Vol. 4851, SPIE Conference Series, 
Truemper J. E., Tananbaum H. D., eds., pp. 28—44 
George 1. M., Eabian A. C., 1991, MNRAS, 249, 352 
Grant C. E., Bautz M. W., ChaiJas G., Gannire G. R, 2004, ApJ, 610, 686 
Guainazzi M., Bianchi S., Dovciak M., 2006, Astronomische Nachrichten, 
327, 1032 

Harrison F A. et al., 2013, ApJ, 770, 103 
Iwasawa K. et al., 2012, A&A, 537, A86 

Kalberla P. M. W., Burton W. B., Hartmann D., Amal E. M., Bajaja E., 
Mon-as R., Pbppel W. G. L., 2005, A&A, 440, 775 


llwasawa et al.ll20T2l : iFalocco et al.ll2013l) . although typically 
only at low significance. However, these studies did not fo¬ 
cus on lensed quasars, and were not able to take advantage of 
the magnification of the intrinsic AGN emission. Addition¬ 
ally, this lensing may preferentially enhance the iron emis¬ 
sion (iChen et all 120121) . further aiding detection. Our anal¬ 
ysis implies that relativistic reflection from the inner disk 
may be relatively common in this lensed population. This 
is a key point, as it demonstrates that with further observa¬ 
tions this population genuinely provides an opportunity to ex¬ 
tend the sample of AGN spin measurements out to high red- 
shift with curre nt instrumentation, beyond the two cas es pre¬ 
sented to date (iReis et al.ll2014l: iRevnolds et al.llMTdl) . Such 
measurements would allow us to begin providing observa¬ 
tional tests of the basic coherent vs chaotic growth scenarios, 
and ultimately of the more s ophisticated models for SMBH 
growth and galaxy formation dPubois, Volonteri & Silkll2014t 
iSesana et aljl2014l) . 


ACKNOWLEDGEMENTS 

The authors would like to thank the reviewer for providing 
feedback which helped improve this paper, and Julian Merten 
for useful discussions. The work of DJW/DS was performed 
at JPF/Caltech, under contract with NASA. 

Facilites: Chandra 


Kara E. et al., 2015, MNRAS, 446, 737 

King A. R., Pringle J. E., 2006, MNRAS, 373, L90 

LaorA., 1991, ApJ, 376,90 

MacLeod C. L. et al., 2015, ArXiv 1501.07533 

Miller J. M., 2007, ARA&A, 45, 441 

Miller L., Turner T. J., Reeves J. N., 2009, MNRAS, 399, L69 

Morgan C. W. et al., 2012, ApJ, 756, 52 

Mosquera A. M., Kochanek C. S., Chen B., Dai X., Blackbume J. A., 
Chartas G., 2013, ApJ, 769, 53 

Nandra K., George I. M., Mushotzky R. F., Turner T. J., Yaqooh T., 1997, 
ApJ, 477, 602 

Nandra K., O’Neill P. M., George 1. M., Reeves J. N., 2007, MNRAS, 382, 
194 

Ota N. et al., 2006, ApJ, 647, 215 
Ota N. et al., 2012, ApJ, 758, 26 

Page K. L., Reeves J. N., O’Brien P. T., Turner M. J. L., Woirall D. M., 
2004, MNRAS, 353, 133 
Reis R. C., Miller J. M., 2013, ApJ, 769, L7 

Reis R. C., Reynolds M. T, Miller J. M., Walton D. J., 2014, Nat, 507, 207 
Reynolds C. S., 2013, Space Sci. Rev. 

Reynolds M. T, Walton D. J., Miller J. M., Reis R. C., 2014, ApJ, 792, L19 
Richards G. T. et al., 2006, AJ, 131, 2766 
Risaliti G. et al., 2013, Nat, 494, 449 

Sesana A., Barausse E., Dotti M., Rossi E. M., 2014, ArXiv e-prints 
Soltan A., 1982, MNRAS, 200, 115 
Thome K. S., 1974, ApJ, 191, 507 

Volonteri M., Sikora M., Lasota J.-P, Merloni A., 2013, ApJ, 775, 94 
Walton D. J., Nardini E., Fabian A. C., Gallo L. C., Reis R. C., 2013, 
MNRAS, 428, 2901 

Walton D. J., Reis R. C., Cackett E. M., Fabian A. C., Miller J. M., 2012, 
MNRAS, 422, 2510 
Walton D. J. et al., 2014, ApJ, 788, 76 

Weisskopf M. C., Brinkman B., Canizares C., Gannire G., Munay S., Van 
Speybroeck L. R, 2002, PASP, 114, 1 
Wilkins D. R., Fabian A. C., 2012, MNRAS, 424, 1284 
Zoghbi A., Fabian A. C., Reynolds C. S., Cackett E. M., 2012, MNRAS, 
422, 129