10/28/2010 SEAL@GSFC 1

MHD Accretion-Disk Winds as AGN X-ray Absorbers

~ Seyfert galaxies to quasars ~

Demos Kazanas Keigo Fukumura Astrophysics Science Division Code 663, NASA/GSFC

Ehud Behar (Technion, Israel)John Contopoulos (Academy of Athens, Greece)

ApJ (2010), 715, 636 ApJ (2010), 723, L228

Credit: NASA/CXC

The Scientific Method

It is a capital mistake to theorize before one has the data. Insensibly, one begins to twist facts to suit theories, instead of theories to suit facts.

Sir Arthur Conan Doyle

It is also a good rule not to put too much confidence in experimental results, until they have been confirmed by theory.

Sir Arthur Eddington

First you get your facts; then you can distort them at your leisure.

Mark Twain10/28/2010 SEAL@GSFC 4

10/28/2010 SEAL@GSFC 5

Some Facts Absorption features are ubiquitous in the spectra of AGN,

GBHC.

50% of all AGN were shown to exhibit UV and X-ray absorption features (Crenshaw, Kraemmer, George 2002).

These features have a very broad range of velocities both in UV and X-rays (a few 100’s – 30,000 km/sec in the UV and a few 100’s – >100,000 km/sec in X-rays).

X-ray features span a factor of ~105 in ionization parameter indicating the presence of ions ranging from highly ionized (H-He - like Fe) to neutral, all in 1.5 decades in X-ray energy!

These “live” in very different regions of ionization parameter space and likely in different regions of real space.

10/28/2010 SEAL@GSFC 6

BAL QSO: X-ray Absorptions

X-ray Absorption line (Fe XXV) Spectral index vs. wind velocity

Brandt+(09); Chartas+(09)

Chandra/XMM/Suzaku

Effect of ionizing spectrum(!?)Fe resonance transitions

X-ray absorber

High-velocity outflows: v/c~0.1-0.7 in Fe XXV/XXVI

10/28/2010 SEAL@GSFC 7

Galactic Black Hole (GBH) BinariesGRO J1655-40:• High ionization: log(x[erg cm s-1]) ~ 4.5 - 5.4

• Small radii: log (r[cm]) ~ 9.0 - 9.4• High density: log(n[cm-3]) ~ 14

Chandra DataMiller+(08)

• M(BH)~7Msun• M(2nd)~2.3Msun

NASA/CXC/A.Hobart

Miller+(06)

Boroson 2002

10/28/2010 SEAL@GSFC 9

Our thesis (and hope) is that these diverse data (including those of galactic X-ray sources) can be systematized witha small number of parameters (2) (Elvis 2000)

Flows (accretion or winds) and their ionization structure are invariant (independent of the mass of gravitating object;ADAF) if:

1. Mass flux is expressed in terms of Eddington mass flux

2. The radius in terms of the Schwarzschild radius

3. The velocities are Keplerian

10/28/2010 SEAL@GSFC 10

10/28/2010 SEAL@GSFC 16

Holczer+(07)

ionization

Behar(09)

AMD( )x = dNH / dlogx ~ (logx)p

column

column

ionization

Absorption Measure Distribution (AMD)

(5 AGNs)

presence of nearly equal NH over ~4 decades in x (p~0.02)

where x = L/(n r2)

(0.02 < p < 0.29)

For radiatively driven winds one obtains

10/28/2010 SEAL@GSFC 17

10/28/2010 SEAL@GSFC 19

Accretion disks necessarily produce outflows/winds (launched initially with Keplerian rotation) Driven by some acceleration mechanism(s) Local X-rays heat up and photoionize plasma along the way

Need to consider mutual interactions between ions & radiation

To AMD through MHD Winds

Blandford+Payne(82)Contopoulos+Lovelace(94)Konigl+Kartje(94)Contopoulos(95)Murray+(95;98)Blandford+Begelman(99)Proga+Kallman(04)Everett(05)Schurch+Done(07,08)Sim+(08;10) & more…

Konigl+Kartje(94)

accelerated

10/28/2010 SEAL@GSFC 21

MHD Disk-Wind Solutions

Steady-state, axisymmetric MHD solutions (2.5D):

5 “conserved” quantities: Energy, Ang.Mom., Flux, Ent., Rot.

(Contopoulos+Lovelace94)

(Prad=0)

Look for solutions that the variables separate

Assume Power Law radial dependence for all variables

10/28/2010 SEAL@GSFC 22

Solve for their angular dependence using the force balance equation in the q-direction (Grad-Safranov equation).

This is a wind-type equation that has to pass through the appropriate critical points.

• With the above scalings

• In order that n(r)~1/r, s = 1 and

• The mass flux in the wind increases with distance!! Or rather, most of the accreting gas “peels-off” to allow only a small fraction to accrete onto the black hole (Blandford & Begelman 1999).

• There is mounting observational evidence that the mass flux in the wind is much higher than that needed to power the AGN/LMXRB.

• Feedback! Edot ~ mdot v2 ~r-1/2 ; Momentum input: Pdot ~ mdot v ~ logr Equal momentum per decade of radius! (talk by J. Ostriker)

10/28/2010 SEAL@GSFC 23

• By expressing BH luminosity in terms of dimensionless variables ( or ) the ionization parameter can now be expressed in the dimensionless variables

• For s=1, x(r) ~ 1/r ; species “living” in lower -x space should come from larger distances.

• The radiation seen by gas at larger distances requires radiative transfer thru the wind (see Keigo Fukumura’s talk).

10/28/2010 SEAL@GSFC 24

The density has a very steep -q dependence with the polar column being 103 – 104 smaller than the equatorial. The wind IS the unification torus (Konigl & Kartje 1994).

10/28/2010 SEAL@GSFC 26

MHD Wind Angular Density Profile

T. Fischer(yesterdy)

e ( - /2q p )/0.2

10/28/2010 SEAL@GSFC 27

Simple Wind Solutions with n~1/r

Density

Launch site

Assume:M(BH) = 106 Msun, G ~ 2 (single power-law), LX ~ 1042 erg/s, mdot ~ 0.5, rad. eff. ~ 10%, n(in) ~ 1010 cm-3

(Fukumura+10a)

Poloidalvelocity

Toroidalvelocity

[cm-3]

(q=1)

10/28/2010 SEAL@GSFC 28

10/28/2010 SEAL@GSFC 29

(Note that different LoS can see different continua; X-ray and UV absorbers need not be identical)

Microlensing technique (e.g. Morgan+08; Chartas+09a) UV regions > X-ray regions (x ~10)

Dust reprocessing: For n(r)~1/r, equal energy per decade of radius is absorbed and emitted as dust IR emission at progressively decreasing temperature. This leads to a flat nuFnu IR spectrum

10/28/2010 SEAL@GSFC 30

Conclusions, Tests

• All accreting BH have winds with velocities reaching v~0.5 c! We do not perceive them because they are highly ionized.

• While NH and x are scale invariant for MHD flows the invariance is broken by atomic physics and radiative processes: – Gas densities scale like ~1/M implying that forbidden

lines should not be present in the galactic LMXRB spectra.

– Low ionization species are also absent in LMXRBs because the size of winds is limited by the presence of the companion.

– The dust sublimation distance in LMXRBs is generally beyond the edge of the disk No Sy2-like IR spectra for galactic sources.10/28/2010 SEAL@GSFC 35

10/28/2010 SEAL@GSFC 36

Apply the model to BAL QSOs by changing only aOX : The decrease in ionizing X-rays allow for FeXXV very close to the BH hi FeXXV velocity, absorption of CIV forming photons CIV forms also at small distances leading to hi CIV velocity (but smaller than that of FeXXV).

mdot = 0.5 kT(in) = 5eV GX = 2 aOX = -2

Injected SED (Fn)Log(density)

Fukumura+(10b)

38

Correlations with Outflow Velocity

Velocity Dependence on SED (X-ray)

10/28/2010 SEAL@GSFC X-ray data of APM 08279+5255 from Chartas+(09) Model from Fukumura+(10b)

Conclusions (final)• We have produced an MHD model for ionized AGN

outflows with 3 parameters: mdot (dimensionless), inclination angle q, and aOX . However the relation between and aOX and luminosity

10/28/2010 SEAL@GSFC 39

Implies that AGN winds can be described with only two parameters . So there is hope for understanding them!

10/28/2010 SEAL@GSFC 40

Summary

We propose a simplistic (self-similar) MHD disk-wind model:

Key ingredients mdot (column) LOS angle (velocity) Fn (SED; G, aOX, MCD…etc.) q (field geometry + density structure)

This model (in part) can account for interesting observables:

Observed AMD (i.e. local column distribution NH as a function of x)

Observed wind kinematics and outflow geometry: Seyferts ~100-300 km/s (Fe XVII); ~1,000-3,000 km/s (Fe XXV) BAL QSOs ~ 0.04-0.1c (UV C IV); ~ 0.4-0.8c (X-ray Fe XXV)

END

10/28/2010 SEAL@GSFC 41

10/28/2010 SEAL@GSFC 42

*Acceleration Process(es)

1. Compton-heated wind (e.g. Begelman+83, Woods+96)

“ Central EUV/X-ray heating a disk thermal-wind” Issue Too large radii…

2. Radiatively-driven (line-driven) wind (e.g. Proga+00, Proga+Kallman04) “UV radiation pressure accelerate plasma” Issues Overionization @ smaller radii… Ionization state freezing out…

3. Magnetocentrifugally-driven wind “Large-scale B-field accelerate plasma” Issue Unknown field geometry…

10/28/2010 SEAL@GSFC 43

Issues (Future Work)

Wind Solutions (Plasma Field):

(Special) Relativistic wind Radiative pressure (e.g.

Proga+00;Everett05;Proga+Kallman04)

Radiation (Photon Field):

Realistic SED (particularly for BAL quasars) Different LoS between UV and X-ray (i.e. RUV > RX by x10…) Including scattering/reflection (need 2D radiative transfer) (Ultimate) Goals:

Comprehensive understanding of ionized absorbers within a single

framework (i.e. disk-wind) AGNs/Seyferts/BAL/non-BAL QSO with high-velocity outflows (e.g. PG 1115+080, H 1413+117, PDS 456 and more…) Energy budget between radiation and kinetic energy…

10/28/2010 SEAL@GSFC 44

Broad Absorption Line (BAL) QSOs

Became known with ROSAT/ASCA survey Large C IV EW(absorb) ~ 20-50 A ~ 30,000 km/sec

~10% of optically-selected QSOs

Faint (soft) X-ray relative to O/UV continua

High-velocity/near-relativistic outflows: v/c ~ 0.04 - 0.1 (e.g. UV C IV) v/c ~ 0.1 - 0.8 (e.g. X-ray Fe XXV)

High intrinsic column of ~ 1022 cm-2 (UV) >~ 1023 cm-2 (X-ray)

10/28/2010 SEAL@GSFC 45

Review on Absorption Features:

• Crenshaw, Kraemer & George 2003, ARAA, 41, 117 (Seyferts)

• Brandt et al. 2009, arXiv:0909.0958 (Bright Quasars)

10/28/2010 SEAL@GSFC 46

END

Chandra surveyGallagher+(06)

Para. CL94 BP82

B r-1 r-5/4

v r-1/2 r-1/2

r r-1 r-3/2

M_wind r1/2 r1/2

Mdot(mass loss rate) ~ 10-6 Mo/yr Fo (ai/1AU)5/2 (M/Mo)-1/2 (Bo/1G)2 ~ 6x1019 g/sec Fo (ai/1AU)5/2 (M/Mo)-1/2 (Bo/1G)2 ~ 6x1013-16 g/sec (ai/1AU)5/2 for M=108Mo, Fo=0.0-0.1, Bo=1-10G

r ~ ai-1 Fo(Bo/vo)2 *r

4710/28/2010 SEAL@GSFC

10/28/2010 SEAL@GSFC 48

Ly a C IV

10. Normal galaxies vs. BAL quasars

Mg II HaHb

Lya NVSi IV

C IV

broad absorption lines(P Cygni profiles)

normal BAL

10/28/2010 SEAL@GSFC 49

Ramirez(08)

10/28/2010 SEAL@GSFC 50

10/28/2010 SEAL@GSFC 51

12. Quasars – SED (UV/X-ray property)

Chandra BAL QSO surveyGallagher+(06)

UV-bright, X-ray-faint!

aox = 0.384 log (f2 keV / f2500 Å)

tells you X-ray weakness

2 keV2500 Å

Elvis+(94)Richards+(06)

?

10/28/2010 SEAL@GSFC 52

Side NotesMCG 6-30-15

Holczer+10

Crenshaw+03

Netzer+(03)

X-ray spectrum of NGC 3783

10/28/2010 SEAL@GSFC 53

fainter in X-rays

228 SDSS Quasars with ROSAT (Strateva et al. 2005)

UV Luminosity vs. aox

brighter in X-rays

Define:Daox = aox - aox (Luv)

log(Luv) (ergs s-1 Hz-1)

10/28/2010 SEAL@GSFC 54

XMM-Newton dataChartas+(09)

G ~ 1.7 – 2.1 Log(NH) ~ 23-24T(var) ~ 3.3 days (~10 rg)

X-ray

10/28/2010 SEAL@GSFC 55Feb. 2010 @Japan

Face-down view (e.g. ~30deg) low NH, low v/c

Optimal view (e.g. ~50deg) high NH, high v/c

(ii) Velocity dependence on LoS

10/28/2010 SEAL@GSFC 56

10/28/2010 SEAL@GSFC 57

10/28/2010 SEAL@GSFC 58

10/28/2010 SEAL@GSFC 59