Collaborators (“UFO Hunters”): M. Dadina, M.Giustini, G. Palumbo, G. Ponti, J. Reeves, T. Yaqoob, V. Braito (but N.1 is Tombesi)

Massimo Cappi and Francesco Tombesi

INAF/IASF-Bologna

X-raying UFOs in AGNs

1. Framework (Multiwavelength) Evidence for UFOs

2. X-ray Observations:A variable phenomenon…up to nowNot really clear…up to now

3. Significance and consequencesInterstellar and intergalactic feedback…

E.T….Mars Attack

(Secret) Outline

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

…known/seen in AGNs since long ago

M87 - JetJets in radio-loud AGNs

Framework: Fast winds/outflows/ejecta in AGNs

Tadhunter & Tsvetanov, Nature, 1989; Wilson & Tsvetanov, 1994Cappi et al. 1995

Wide-angle winds & jets in Seyfert galaxies

Sey2 NGC5252OIII cones

Fast (v up to ~ 50000 km/s) winds in BAL QSOs (~ 20% of all QSOs)

Weymann et al., ’91; Reichards et al., ‘03

See yesterday’s talks by Kriss and Gallagher

Fabian, et al. ’94Otani, ’95, PhDReynolds et al. '97Georges et al. '97

50% of all Sey 1s exhibit WAsASCA

Many details from Chandra/XMM gratingsNGC3783 Exp=900 ks

Kaspi et al. '01; Netzer et al. '02; Georges et al. '03; Krongold et al. ‘03

Clear now that often multiple ionization & kinetic components: outflows with v~100-1000 km/s

Blustin et al. 2004

Framework: Warm absorbers…probe highest-v outflowing/ejected gas

Framework: Blue-shifted absorption lines/edges – High-v

PG1211+143 (z=0.08) v~0.1c

2 Energy (keV) 5 7 10

New and unexpected results from Chandra and XMM-Newton observations

Pounds et al. 2003a,b

massive, high velocity and highly ionized outflows in several RQ AGNs/QSOsMass outflow rate: comparable to Edd. Acc. rate (~M◉/yr); velocity ~0.1-0.2 c

(If) interpreted as Kα resonant

absorption by Fe XXV (6.70 keV)

or FeXXVI (6.96 keV)Reeves et al. 2003

PDS456 (z=0.18) v~0.1c

2 Energy (keV) 5 7 10

Absorbers variability on timescales 1000-10000s Framework: Blue-shifted absorption lines/edges - Variability

Risaliti et al. 2005

NGC1365 Mrk 509 (long-look, 200ks)

MC et al., 2009Dadina et al. ‘05

Obs1

Obs2

Obs3

Variability allows to place limits on location, mass, etc.(See also Krongold et al. 2007 on NGC4051)

MRK509 contour plots (MC et al. 2009)

Mrk 509: Among brightest F~2-5x10-11 cgsand most luminous L~1-3x1044 ergs/sType 1 Sey known

EW(FeXXVI)~-20 -60 eVV~0.14-0.2c∆t~100 ks

Log~5 erg cm/s, Nh~2-4x1023 cm-2,

v~0.14-0.2c

From absorber parameters and variability, we estimated:R<500 Rs

MC et al. 2009

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.QuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

Again, absorber variability on timescales ~20000s MCG-5-23-16 (XMM+Chandra)

Braito et al., 2007

X-ray Observations: Variability

Data require large Nw, high , and vout=0.1c

iii) Magnetically driven winds from accretion disk

Emmering, Blandford & Shlosman, ’92; Kato et al. ‘03

Interpretation: (Three main) Wind dynamical models

i) Thermally driven winds from BLR or torus

Balsara & Krolik, 93; Woods et al. ‘96

i) Large R, low vii) and iii) Low R and large v

Murray et al. ‘95, Proga et al. ‘00

ii) Radiative-driven wind from accretion disk

…and/or…

Theoretical Interpretation: (Three main) Wind dynamical models

MHD+LD model by Proga et al. ‘00, ‘03

QuickTime™ and aYUV420 codec decompressor

are needed to see this picture.

Overall numbers (Nh, , vout, etc.) are consistent with observations…

Data Interpretation:

Sim et al., 2008

Yes indeed…one expects (mostly/only) strong Fe line absorptionswhen accounting for proper wind geometries and physics

Most important (open) issue

Nw (cm-2)

Location (R, DeltaR)

Ionization state ()

Velocity

Covering factor

Frequency in AGNs

Current estimates have order of magnitude uncertainties, they go from:

dM/dt (Lkin) few % to several times dMacc/dt (Ledd)

This is a fundamental (open) issue

Elvis et al. ‘00, Creenshaw et al. ’03, King et al. ‘03, Chartas et al. ‘03,

Yaqoob et al. ‘05, Blustin et al. ‘05, Risaliti et al. ’05, Krongold et al. ‘07

Fundamental to:

i) PHYSICS of accelerated and accreted flows (winds?, blobs?, etc.), i.e. understand how BHs accelerate earth-like quantities of gas to relativistic velocities

ii) COSMOLOGY: i.e. estimate the mass outflow rate, thus the impact of AGN outflows on ISM and IGM enrichment and heating!

• Selection of all NLSy1, Sy1 and Sy2 in RXTE All-Sky Slew Survey Catalog (XSS; Revnivtsev et al. 2004)

• Cross-correlation with XMM-Newton Accepted Targets Catalog

• 44 objects for 104 pointed XMM-Newton observations

• Local (z<0.1)

• X-ray bright (F4-10keV=10-12-10-10 erg s-1 cm-2)

z distribution of sources 4-10keV fluxes

We analysed in a systematic and uniform way, a (almost) complete sample of nearby, X-ray bright, radio-quiet AGNs (Tombesi et al. 2010a,b)

Ejection/outflows: From source-by-sources to representative samples…

Absorption lines search

Uniform spectral analysis:• Reduction and analysis of all EPIC pn spectra in the 4-10keV

• Baseline model: absorbed power-law + Gaussian Fe K emission lines

Example of PG1211+143 (Tombesi et al. 2010a)

Absorption lines search:• Addition of narrow line to baseline model stepping energy in 4-10keV and recording Δχ2 deviations

• Visualization on energy-intensity contour plot (significance 68% red, 90% green, 99% blue) (e.g. Miniutti et al. 2007, MC et al. 2009)

• Selection of narrow lines with F-test confidence levels ≥99%

• Line parameters determined by direct fitting to the data

Absorbed power-law

Absorbed power-law + emission line

Absorption lines significance

Extensive Monte Carlo simulations (e.g. MC et al. 2009)

• Additional significance test for lines at energies ≥7.1keV

• Null hypothesis that spectra are fitted by model without absorption lines

• 103 simulated spectra for each case

• Simulated Δχ2 distribution for random generated lines

• Selection of lines with MC confidence levels ≥95%

F-test can overestimate the detection significance for a blind search of emission/absorption lines over a range of energies (e.g. Protassov et al. 2002).

Global probability for the lines to be generated by random fluctuations is very low (≤10-8 from Binomial distribution).

• Checked no contamination from pn background and calibration

• Independent confirmation of blue-shifted lines detection from MOS data (without relying on any statistical method)

Results: Yes, we confirm there are indeed UFOs! (Ultra-Fast Outflows…)

• 36 absorption lines detected in all 104 XMM observations

• Identified with FeXXV and FeXXVI K-shell resonant absorption

• 19/44 objects with absorption lines (≈43%)

• 17/44 objects with blue-shifted absorption lines (lower limit ≈39%, can reach a maximum of ≈60%)

• 11/44 objects with outflow velocity >0.1c (≈25%)

• Blue-shift velocity distribution ~0-0.3c, peak ~0.1c

• Average outflow velocity 0.110±0.004 c

Blue-shift velocity distribution Cumulative velocity distribution

Tombesi et al. 2010aQuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

(The UFO’s hunters commander in chief)

Results

• Most frequent detected line is FeXXVI Lyα

• EW is in the range ≈10-100eV, with mean ≈40-50eV

• Estimated global covering factor from fraction of sources with lines (C=Ω⁄4Π)≈0.4-0.6

• Geometry not very collimated, large opening angles favored (similar to WA)

Lines EW distribution

Physical photo-ionization modelling using XSTAR

Xstar grid for direct spectral fitting:

• input enenergy band 0.1eV, 106 eV

• mean SED for radio-quiet sources

• power-law Γ=2 for radio-loud sources

• turbulent velocity v=500 km/s

Mean phenomenological SED from the radio-quiet sample:

• SEDs 34 Type 1s from NED database

• scattering due to source variability and different instruments

• normalized SEDs to near-IR inflection point ~1.25μm (like Elvis et al. 1994)

• average flux for each energy point

Simple mean SED with three intervals: Radio to mm (Γ~2), mm to IR (Γ~0.1), IR to X-rays

(Γ~2)

NH~1022-1024 cm-2, logξ~3-6 erg s-1 cm-2

(vout, NH lower limits, unknown inclination angle)

WAs in RQ from McKernan et al. (2007) (filled black circles), WA in RL from Torresi et al. (2009) and Reeves et al. (2009) (filled blue triangle), UFOs in RQ (red crosses)

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Tombesi et al., 2010c, in prep

Ejection/outflows:• estimated distances: r<0.01-0.1pc (<102-105 rs)

(accretion disk winds? e.g. Elvis 2000; King & Pounds 2003)

• Often vout > vesc, but not always, material shall fall back sometimes? (“aborted jet”? e.g. Ghisellini et al. 2004, Dadina et al. 2005)

• variability time scales t~1day – 1year

• Lbol/LEdd~0.1-1; Mout/Macc~0.1-1; Ek~1044-1045 erg s-1 ~0.1 Lbol

(last two estimates depend on covering fraction C)

• Acceleration mechanism? Line, magnetically or momentum driven?

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Magorrian et al. '98Tremaine '02; Gebhardt '02...etc

Cosmological Importance of UFOs?: (on-going work)

Mbh~ б4

Role in feedback in the (co?)evolution of galaxies?

Lapi, Cavaliere & Menci, ‘05

Grav. scalingWith SN preheating

With AGN pre-heating

With QSO ejection/outflows

(see e.g. King and Pounds '03, Crenshaw, Kraemer & George '03, ARA&A)

Role in heating groups and custers?

Impact on physics of ejections/outflows?

Momentum-driven accretion disk winds/outflows? As in King 2009?

“Eddington winds from AGN are likely to have velocities ~0.1c and show the presence of helium- or

hydrogen-like iron” (King 2009)

• Eddington accretion episodes Lbol/LEdd~0.1-1

• electron scattering wind τ~1 at infinity

• wind momentuum ~ photon momentuum (Moutv~LEdd/c)

• typical velocity ~0.1c

• typical ionization parameter logξ~4 and linear relation vout and ξ

(Fe XXV, Fe XXVI + Nickel ions?)

• wind interaction with host galaxy, ram pressure

• important for feedback SMBH and host galaxy

• does explain the M-σ relation

Future prospects: Astro-H calorimeter

Calorimeter on board Astro-H, next Japanese X-ray satellite to be launched in 2013: good effective area (≈250 cm2 @ 6keV) and high energy resolution (FWHM≈7eV) from 0.1keV up to 12-13keV.

• realistic spectra simulations of UFOs

•absorption lines resolved, measured velocity broadening (down to ~100-200 km/s @ 6 keV)

• estimates EW, blue-shift, centroid energies ~10 times better XIS-FI (100ks simulation)

• better constrains on NH, logξ, vturbu

Future pospects: IXO calorimeterX-ray Microcalorimeter Spectrometer (XMS): high effective area ~0.65m2 (~6500cm2 !!) @ 6keV and high energy resolution (FWHM≈2.5eV) from 0.1keV up to 12-13keV.

Flux limits• 2-10keV flux limits for 5σ detection of narrow absorption lines in the 3-11keV

• Different EWs, exposure times and responses

• Lines of EW=10eV (50eV) in ≈6-9keV for ≈10-12

(10-13) erg s-1 cm-2 (expo 100ks)

• Spectral variability on time-scales of 5 (10) ks for ≈10-11 (10-12) erg s-1 cm-2

Flux limits (EW=10eV) (Tombesi et al. 2009) logξ=3 erg s-1 cm, NH=1023cm-2, b=1000km/s (Tombesi et al. 2009)

Spectra simulations

• Simulations of highly ionized and massive absorbers

• FeXXV/XXVI K lines detectable with high significance

• Line details (profile, energy, broadening) measured with high accuracy (>30 times Astro-H)

• Extend study to less bright sources

• Time variability, dynamics of absorbers

NGC1365 F(2-10)=10-11cgs S/N>3

Highest throughput for time-resolved detections of abs. lines real-time, extreme dynamics, i.e. inward and outward accelerations!?(line ∆v/∆t) ....blob=test particle to test Kerr vs. Schwarzschild GR

Future: Time vs. energy maps (in emission and absorption lines)

Simula

tion

Credit:F

. Tom

besi

Accelerated flow Decelerated flow

Conclusions

• Search for narrow blue-shifted Fe K absorption lines in a complete sample of 44 radio-quiet AGNs observed with XMM-Newton

• 36 detected absorption lines (22 at E≥7.1keV)

• Global veracity is strong and publication bias solved

Existence of highly ionized, massive and ultra-fast outflows in radio-quiet AGNs:

• ≈40% of sources have blue-shifted absorption lines (≈25% with v≥0.1c)

• Outflow velocities up to relativistic values (≈0.2-0.3c)

• Global covering factor ≈0.4-0.6, large opening angles favored

• Important for: BH accretion physics, AGN feedback with host galaxy, SMBH growth, ...

• Improvement expected from future X-ray missions, such as Astro-H and IXO

Thanks for your attention, and watch-out for UFOs!

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

and don’t be confused by the fact that UFOs are also called “Ultra-Fast Outflows”…

APM 08279+5255 (z=3.91) v~0.2-0.4c

2 high-z BAL QSOsChartas et al. 2002, Hasinger, Schartel & Komossa 2002

N.B.: Would have been undetected at z=0...

Massive outflows…also (mostly?) at high redshift

X-ray Observations (ii/iii): Blue-shifted absorption lines/edges – High-z

See also Wang et al. ’05 (v=0.8c in qso@z=2.6)Chartas, Brandt & Gallagher, 2003

PG 1115+080 (z=1.72) v~0.1-0.3c

MOS

PN

Critical Issues (i/ii): Observations

Lines statistical significance? (transient features, number of trials in time and energy, etc…)

Identifications of edges/lines energies? (Kallman et al. 2005, Kaspi et al. for PG1211)

Local “contamination”? (PDS456 at risk? McKernan et al. ’04, ‘05)

Publication bias? Only positive detection, low signif., Vaughan & Uttley ‘08)

cz (km/s)

Out

flow

v (k

m/s

)

Last but not least…for those still skepticals on UFOs

Publication bias solved :

• Uniform analysis on complete sample of sources

• Lines detection assessed by MC simulations

• Global chance probability very low (≤10-8)

• Detection independently confirmed by MOS data

Not due to local contamination:

No correlation between cosmological red-shift and lines blue-shift, no local (z≈0) absorption.

See poster E17-0049-08 by Giustini et al.

Future (iv/vii): Better statistics on optically-classified samples:

Giustini et al., submitted

SDSS’s BALQSOs cross-correlated with 2XMM 22 spectra and 23 HRs to estimate Nh, , Lx, etc.

NGC1365 F(2-10)=10-11cgs S/N>3Future (v/vii): Shorter time-scales and better sensitivity…with XEUS

XMM resul

ts

Risaliti

et al

. 200

5

XEUS S

imula

tion

Area (cm2) @1 keV @4 keV @6 keV @7 keV

XEUS 58000 40000 30000 20000IXO 30000 15000 9000 6000CON-X 15000 11000 7500 5000XMM(PN) 1150 820 800 650

Aeff Flength2 x crit2 x Refl

2 , where c 0.5/E

Future (vi/vii): Con-X, XEUS or IXO?

SummaryI briefly reviewed the current evidence for blue-shifted

absorption lines from highly ionized Fe in AGNs (in both Sey and QSOs)

These indicate the existence of highly-ionized, high velocity, massive outflows in AGNs, BUT STILL ORDER OF MAGNITUDES UNCERTAINTIES

on energy/momentum and mass involved.

This topic still requires better measurements of intensity, energy and frequency/recurrency but has a great potential for the study of:

launching mechanisms/characteristics of outflows/jets (mechanical energy emerging from BH), important not only for

(relativistic) physics but also for link with cosmology

Prospects for future progress are to:i) Confirm/secure these findings with longer XMM/Chandra observationsii) Probe lower time-scales (with > 2m**2 @6 keV, i.e. XEUS-like mission)iii) Probe high-velocity gas with high-energies (with Simbol-X and/or XEUS

hxd)

…to probe the flow dynamics (∆v/∆t) of innermost regions by means of detection and time-resolved spectroscopy

of energy-shifted absorption lines.

Fiducial numbers:We wish to follow abs. lines from, say, ~1 to ~10 Rs, with intervals of 1RsLet assume v~0.2c, then for BH mass= 108 M● ∆Time-scale ~ 5000 sBH mass= 106 M● ΔTime-scale ~ 50 s (Note: 1µs if 1 M●)

Scaling from Mrk509 and XMM, and assuming EW(Fe)=-100 eV

Con-X(6000cm2) XEUS(25000cm2)@6keVF(2-10)=2x10-11cgs (~ 15 sources) 1000s 100sF(2-10)=2x10-12cgs (~ 50 sources) 10000s 1000sF(2-10)=2x10-13cgs (~ 250 sources) 100000s 10000s

N.B.: If v>0.2c, timescales consequently reduced

Really needed because mostly BH mass≤ 107 M●

Future (iii/vii): Reduce timescales…

X-ray spectra of winds/outflows

Model with narrow emission and absorption lines: PL (Г=1.9, F(2-10)=10-11erg/cm2s, Exp.=50 ks) + 2 FeK emission lines (E1=5 keV, E2=6.4 keV, σ1= σ2=50 eV, EW1=EW2=100eV) + 4 FeK absorption lines (E1,2,3,4=7, 9, 12, 15 keV, σ1,2,3,4<50 eV, EW1,2,3,4=-100eV)Edges and absorption lines at E~7.1-9.0 keV

(rest-frame) + vout~ 0.1-0.5c Eobserved~ 8-14 keV !!

Simulations of narrow emission and absorption lines

Simula

tions

Model: PL + 2 emission lines + 4 abs. lines

Eabs<100 eV Idee would be to follow the evolution of blob ejections (or injections) N.B: Masses involved can be greater than Mearth (1027g/ejecta) >>10-11g in accelerators

Deceleration Acceleration

Future (vi/vii): Simbol-X simulation (SDD+CdTe)

Mrk 509, Astro-E2 simulation 100 ks

Unfortunately XRS on-board ASTRO-E2 is lost

High energy resolution to distinguish beetwen wind and blob(s) (line profile)

Simula

tions

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.QuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.