Solving Quasars part I

Solving QuasarsSolving Quasarspart Ipart I

Martin ElvisHarvard-Smithsonian Center for Astrophysics

in particular…

Understanding Quasar Atmospheres

Elvis M., 2000, Astrophysical Journal 545, 63

Fermilab Colloqium 29 October 2003

QuasarsQuasars** unsolved after 40 years unsolved after 40 years

Once were a `hot topic’ Were the first to start the downfall of Steady State Cosmology

- via ‘evolution’: change in density with cosmic time

Now astronomers have moved on to easier problems – Large scale structure, Dark Energy and Gamma-ray bursts

Quasar studies continue to generate many papers …but little understanding?

Discovered in 1963

Quasars are the most powerful continuous radiation sources in the Universe

* Note for the pedantic: By ‘quasars’ I mean all types of ‘activity’ in galaxies


What’s the problem?What’s the problem?

Enormous array of detail

We have no images of a quasar atmosphere

Would need 1000 times sharper pictures than Hubble or Chandra <100as

Must rely on spectra: span all wavelengths: X-ray optical - radio

Superficial understanding


Why Study Quasars?Why Study Quasars?

We live on a planet A star gives us life Galaxies dominate the Universe … but why do quasars matter? Here are 4 answers:


Outside the wavelength range that our eyes are sensitive to

1. An Astronomer’s Answer1. An Astronomer’s Answer

Radio

Gamma-ray

X-ray

Quasars dominate the night sky


2. An Astrophysicist’s Answer2. An Astrophysicist’s Answer

Gravity powered, not fusion. via Black Holes 106 - 109 as massive as the

Sun. Gas heats up falling toward it, like a spacecraft on re-entry.

Billions of times brighter than stars. Can outshine a whole galaxy

Emit strongly from radio to -rays. How do they do that?

Make galaxy length jets


The power available from gravity for heating is all too obvious following the Columbia tragedy

3. A Cosmologist’s Answer3. A Cosmologist’s Answer

Emit up to 1/5 of power in Universe: Important input, may dominate in some places, times.

Quasars lie at the hearts of galaxies:

Galaxy mass and quasar black hole mass are tightly connected. Maggorrian et al, Ferrarese & Merritt, Gephardt et al.

How? Should be governed by different processes.

Already exist at t<1Gyr (z=6)FIRST survey discovery, Becker et al.

½ Gyr from WMAP reionization at z=20

special role in early Universe?

reionization, seeding galaxies…

element creation, star formation catalyst via dust creation?

4. A Physicist’s Answer4. A Physicist’s AnswerEject bulk gas at 99.50% speed of light

=10

similar to proton in Fermilab Tevatron 99.88%c

Impacts gas of intergalactic medium.->what emerges?

Accelerates e- to ~1000 -> TeV photons

X-rays come from region of Strong Gravityseen in 6.4 keV `Fe-K’ (=Fe Lyman-) emission line?

Wilms et al., 2002, MNRAS, 328, L27

6.4 keV

GR redshift?

MCG -6-30-15 (XMM)

Reynolds C.

Reynolds C.


What What dodo we know? we know?High level theory rapidly gave a clear picture

massive black holeLynden-Bell 1969

accretion diskLynden-Bell 1969, Pringle & Rees 1972,Shakura & Sunyaev

1972

relativistic jet Rees 1967 [PhD], Blandford & Rees 1974

All established just 10 years after discovery


does not connect to the atomic physics

features observed in quasars

This theory describes a naked quasarThis theory describes a naked quasar

Leaves us with no way to order observations, nothing to test


Atomic features in Quasar AtmospheresAtomic features in Quasar Atmospheres

High ionization: e.g. CIV, OVI

Low ionization: e.g. MgII, H.


All studied separately with separate telescopes

Compton gamma-ray Observatory

Chandra

Hubble

MMT

Sub-millimeter array

VLA

Quasars have no temperatureQuasars have no temperatureWhipple 10 meter

Blind men and the elephant. Manga VIII Hokusai, Katsushika (1760-1849)

Wall, Tree, Rope, Spear, Snake, Fan

Not having the complete picture can be misleading


we need a ‘low theory’we need a ‘low theory’ that deals with the multitude of quasar details

Just as there are textbooks on ‘Stellar Atmospheres’ we need the subject of ‘Quasar Atmospheres’ Takes more than 1 step. First build an observational paradigmobservational paradigm

i.e. what do the observations drive require of any theory?

These optically thin features are all interconnected

= `quasar atmosphere’


A Paradigm for Quasar AtmospheresA Paradigm for Quasar Atmospheres

Accretion disk

Broad Absorption LinesReflection features

Thin Vertical wind

Narrow absorption lines X-ray `warm’ absorbers

Broad Emission Lines

hollow cone

Supermassive black hole

Elvis M., 2000, Astrophysical Journal 545, 63

X-ray/UV ionizing continuum

Accelerating bi-conical wind

NB: Independent of Unification Jets are not included

no absorption lines

A Geometric & Kinematic solutionc.f. Rees relativistic jets for blazars/radio sources

Can now re-construct this model using data not in Elvis 2000

Quasar AtmosphereQuasar Atmosphere


Take a lesson from lab plasmas: use all the dataTake a lesson from lab plasmas: use all the data

NSTX diagnostic instruments cover everything

NSTX at PPPL

National Spherical Torus eXperiment

Princeton Plasma Physics Laboratory

2mm interferometer

X-ray PHA

X-ray crystal spectrometer

Radiometer

Thomson scattering

Far infrared tangential

Interferometer/polarimeter

Visible spectrometer

Vacuum UV survey spectrometer

Grazing incidence spectrometer

Tangential bolometer array

Single channel visible

Bremsstrahlung detector

Polarimeter

X-ray pinhole camera

Soft X-ray arrays

Fast tangential X-ray camera

Reflectometer array

Infrared cameras

Princeton AGN Physics with the SDSS, 29 July 2003

12,277 Papers on Quasars since 196312,277 Papers on Quasars since 1963** *ADS to 4/18/03, refereed only , search on abstract containing ‘quasar’ | ‘AGN’

1/day. Now 2 per day = 5% of all astronomy papers

SpamSpam!! Need filters---

1. Physical measurementsMass, length, density. Not ratios, column densities

2. Favor absorption: advice from Steve Kahn c.1985

1-D spatial integral, not 3-D;

blueshift = outflow

3. Use Polarization Non-spherical geometry

With these filters just a dozen papers define the structure of quasar atmospheres.


1.Physical Measurements: 1.Physical Measurements: BEL Velocity-radius relationBEL Velocity-radius relation

Reverberation mapping shows Keplerian velocity relation in BELs

Pole-on

Keplerian orbits

Mass

Light echo delay (days) Dopple

r w

idth

of

em

. Li

ne

Peterson & Wandel 2000 ApJ 540, L13

Broad Emission Lines close to Keplerian velocities

~1000 rs,

Schwartzchild

radii


1.Physical Measurements:1.Physical Measurements: AngleAngle

Pole-on

Use VLBI + X-ray to get angle of jet to line of sight Rokaki et al. 2003 astroph/0301405

(2) Continuum drops as cos EW=EW0[1/3 cos(1+2cos)]-1 limb darkened disk

H does not H scale height larger than disklike

optical continuumBut BLR is rotating

rotating cylinder?A highly non-equilibrium shape

Simplest solution: BLR is in a rotating wind

Edge-on

Pole-on

SLOW

FAST

Flat disk continuum

Relativistic beaming

isotropic

Con

tin

uu

m/H

f

lux

Rokaki et al. 2003 astro-ph/0301405

(1) Rotation about jet axisc.f. Wills & Browne 1986, Brotherton

1996, McLure & Jarvis 2003

H polarization rotation also implies orbiting gas Smith et al 2002


2. Absorption Features2. Absorption Features

Narrow UV lines

High ionization CIV, OVI High ionization OVII,OVIII

Outflow ~1000 km s-1

Seen in 50% of quasarsSeen in same 50% of quasars

Simplest solution: Same gas, 2 phases


Winds are common in quasarsChandra HETG: 900ksec NGC3783

new

Narrow X-ray linesnew

Same Outflow ~1000 km s-1 new

Chandra HETGS 850ksec spectrum of NGC 3783

Over 100 absorption features fitted by a 6 parameter model One T~106 K and one T~104 K, in pressure balance to 5%

Krongold, Nicastro, Brickhouse, Elvis, Liedahl & Mathur, 2003 ApJ, in press. astro-ph/0306460

2-phase gas in pressure equilibrium

2. Absorption: More Physics from X-rays2. Absorption: More Physics from X-raysFermilab Colloqium 29 October 2003

2. Absorption: where is the wind?2. Absorption: where is the wind?

Velocity dependent covering factors

Absorber is close to continuum source

absorber is moving transverse

Wind is close to continuum, crosses line of sight

Arav, Korista & de Kool 2002, ApJ 566, 699

Arav, Korista, de Kool, Junkkarinen & Begelman 1999 ApJ 516, 27


A quasar wind is like a flameA quasar wind is like a flame

We are looking through a flow

Apparent lack of change is a common handicap for astronomers

the ‘Static Illusion’

e.g. expansion of the Universe, cluster cooling flows, quasar disks


Emission lines: a thin wind?Emission lines: a thin wind?

Narrow Line Seyfert 1 galaxies (NLSy1s) show

broad, strongly blueshifted

high ionization (CIV) lines Understandable as disk wind redshifted lines hidden by disk Low ionization lines from outer

disk c.f. Collin-Souffrin, Hameury & Joly,1988 A&A 205, 19

Leighly & Moore 2003, ApJ submitted

See: Gaskell 1982

Wilkes 1984

Low ionization

MgII

BELs are rotating, transverse, thin winds


2. Absorption / 1. Physical Measurements:2. Absorption / 1. Physical Measurements:Wind Density,thicknessWind Density,thickness

UV/X-ray absorption responds to continuum changes: photoionized

Nicastro et al. 1999 ApJ, 512, 184

Absorbing wind is dense

But responds with a delay = recombination/ionization time density ne~108 cm- 3 for OVIII

ne~3x107 cm-3 for FeXVII

X-ray continuum

“OVII edge”

“OVIII edge”

time

R

wind

accretion disk

Black hole

R

To Earth

density + column density (~3x1022 cm-2)

thickness (~1015 cm) < distance to continuum

Absorbing wind is narrow


3. Polarization: X-ray absorbers3. Polarization: X-ray absorbers

Absorption line quasars are highly polarized in optical:

1. Scattering off non-spherical distribution

Edge-on structure

2. Pole-on objects must be unobscured

scatterer & obscurer:

flattened & co-axial

Leighly et al. 1997 ApJ 489, L137

Absorbers are seen edge-on


Flattened, Transverse Wind Flattened, Transverse Wind axisymmetryaxisymmetry

A transverse wind suggests an axisymetric geometry:

bi-cones looking edge-on see

absorbers Wind does not hug disk pole-on: no absorbers absorbers in all quasars

Mathur, Elvis & Wilkes 1995 ApJ, 452, 230


Absorbing wind is a bi-cone to 1st order

Putting X-ray/UV absorber and BEL togetherPutting X-ray/UV absorber and BEL together

Both are disk winds rising well above the disk plane

Elvis 2000 ApJ 545, 63; Krongold et al. 2003

Similar Pressure:

P( abs ) ~1015 = 104 K x 1011 cm-3

P(BELR)~1015 = 106 K x 109 cm-3

Matching Ionization Parameter, U:T/U( abs ) = 106 = T( abs ) ~106 K/ U( abs ) ~1

T/U(BELR)= 106 = T( abs ) ~104 K/ U(BELR) 0.04

Similar Radius: for NGC 5548

r( abs ) ~1015 - 1018cm recomb. time + NHX

r(BELR)~1016cm CIV reverberation mapping

Keep it simple: Emission and Absorption are 2 phases of the same quasar wind

Similar Temperatures

For low U absorber, BELsnew

They share physical properties:


Components of Quasar AtmospheresComponents of Quasar Atmospheres



In a 2-phase transverse wind in pressure balance

United


The Final Element:The Final Element:Broad Absorption Lines (BALs)Broad Absorption Lines (BALs)

Old question: Special objects? or Special angle?

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

Ferland & Hamann 1999 Annual Reviews of Astronomy & Astrophysics , 37, 487

10% of quasars show BALs with doppler widths ~2%c - 10%c

~10x NALs. Clear acceleration (or deceleration)


Broad Absorption Lines (BALs)Broad Absorption Lines (BALs)

BEL FWHM correlates with BAL velocity (at minimum flux)

V(BAL) ~ 2 FWHM(BEL)

Lee & Turnshek 1995 ApJ 453 L61


BAL gas knows about BEL gas

More BEL-BAL correlations:

Reichard et al. 2003B

EL

wid

th

BAL width

2:1

BALs from a rotating windBALs from a rotating wind

Redshifted BAL onset Possible occasionally in

a rotation dominated wind

Hall et al. 2002 ApJS, 141, 267

BALs need a rotating wind… like the BELs

BALEm.

line

wavelength

Continuumflu

x

Detachment velocity

blue red


3. Polarization: BAL troughs3. Polarization: BAL troughsOgle et al. 1999 ApJS, 125, 1; Ogle 1998 PhD thesis, CalTech

BAL troughs are highly polarized – scattered light off flattened structure

=> BALs are common. Universal?Scattering solves other BAL problems:

ionization, abundances, NH

Thomson thick: X-ray Fe-K, Compton hump

Ogl

e, P

hD th

esis

, 199

8

Conical wind fits BALs well

Hamann 1998 ApJ 500, 798; Telfer et al. 1998 ApJ 509, 132

Is the BAL wind itself the scatterer?

Bi-cone model Predicts distribution of non-BAL quasar polarization


3. Polarization: VBELR3. Polarization: VBELR

Supported by observations: Emission lines twice as broad in

polarized, non-variable light. non-BAL quasars have

Thomson thick gas at large, BAL, velocities

Don’t see in absorption because out of our line of sight

Large scattering region (but not too large, Smith et al. 2003

MNRAS) with BAL velocities

Young et al. 1999 MNRAS 303, 227

Polarized light2 x width

total light

MKN 509

BAL velocity gas exists in non-BAL quasars

If BALs are cones, all quasars should have BAL gas


One last, crucial, complicationOne last, crucial, complicationAngles are wrong:

BAL velocities too high: ~10,000 km s-1

10 times narrow absorption lines

Requires extreme cone opening angle.

Simple solution: bend wind

Predicts:

1. ‘detached BALs’*

= Lowest velocity where wind bends into our line of sight

= vertical velocity from disk

2. ~10% covering factor

r at r gives 6o divergence angle

radiation forces gas to diverge

Both previously unexplained

BALEm.

line

wavelength

Continuumflu

x

Detachment velocity

*Could this be an ionization effect? v IP?


Quasar Atmospheres, Quasar WindsQuasar Atmospheres, Quasar Winds

60 deg: broad absorption lines

20 deg: no absorption lines

One geometry unites all the features

85 deg: narrow absorption lines

High ionization Broad emission lines Low ionization


Components of Quasar AtmospheresComponents of Quasar Atmospheres



All atomic features now included

Thompson thick BAL scatterer must also make Compton hump, Fe-K


Putting it all togetherPutting it all togetherinformation filters worked efficiently!

Accretion disk

hollow cone

Supermassive black hole

Elvis M., 2000, ApJ, 545, 63

X-ray/UV ionizing continuum

BALs

Polarization

Accelerating bi-conical wind

Thin quasi-vertical wind

no absorption lines

NALs

BELsWAs


Blind men and the elephant. Manga VIII Hokusai, Katsushika (1760-1849)

Hokusai never saw a live ElephantHokusai never saw a live ElephantNot bad – not 100% right – but gets the idea

This picture of quasar atmospheres is probably in much the same state: needs physics bones


A Quasar Observational ParadigmA Quasar Observational Paradigm

Disk Winds: tie together all the pieces of the quasar atmosphere

Explains features not ‘built in’ BAL covering factor; detachment velocity, Hi ionization BEL blueshifts.

Survived tests X-ray absorber outflow v, 2-phase UV/X-ray absorber, pressure balance

Makes predictions High ionization BEL, X/UV absorber radii, thickness are equal

Creates a research program c.f. Lakatos 1980

Allows tractable physics exploration… Work BACK to origin in accretion disk physics Work OUT to impact on surroundings

Can begin to build a ‘low’ theory of quasar atmospheres


low theory: 2-phase equilibriumlow theory: 2-phase equilibriumKrolik, McKee & Tarter 1981, ApJ, 249, 422

•Photoionized gas tends to have phases

• Not really new:

•Does not work for a static medium

•so abandoned…. a mistake!

•Works fine in a wind. dynamic

•Equilibrium determined solely by: SED & ionization thresholds

•Should be similar from object to object

•No need to assume ‘clouds’

Krongold, Nicastro, Brickhouse, Elvis, Liedahl & Mathur, 2003


low theory: accretion disk physics, IIlow theory: accretion disk physics, IIKrongold et al. in preparation

new

•~106K phase depends critically on SED Nicastro 1999, Reynolds & Fabian 1995

• Use absorber (T,x) to determine unseeable EUV SED

-> Test models of accretion disk

•inner edge ill-defined- boundary condition

•‘plunging region’ Krolik et al.

Krongold, Nicastro, Brickhouse, Elvis, Liedahl & Mathur, 2003

Reynolds & Fabian 1995 MNRAS 273 116


low theory: Why is the wind thin?low theory: Why is the wind thin?

Intermediate level 2D theory Wind driven by UV absorption lines

c.f. O-star winds, CAK ignore gas pressure

3 Zones: Inner, Middle, Outer 1. Inner: over-ionized

Only Compton scattering - insufficient shields gas further out from X-rays =

Murray & Chiang `hitchhiking gas’

2. Middle: UV absorption drives gas wind escapes

3. Outer: shielded from UV, weak initial push from local disk radiation – wind falls back

Risaliti & Elvis 2003, ApJ submitted

wind escapes

Outer: wind falls back

Inner:

`failed wind’

Middle

No wind

Wind

new

•Note: L>LEdd quasars always have winds

•See King & Pounds 2003 astro-ph/0305571

•Reeves et al … ;Chartas et al…

Some BALs are L>LEdd windsweakness

density


Looking Out: quasars as dust factoriesLooking Out: quasars as dust factoriesElvis, Marengo & Karovska, 2002 ApJ, 567, L107

Cooling BEL clouds

Oxygen rich dust

Cooling BEL clouds

Carbon rich dust

Outflowing BEL gas expands and cools adiabatically BEL adiabats track through dust formation zone of AGB stars

Applies to Carbon-rich and Oxygen-rich grains


Outflows rates ~10 Msol/yr at

L~1047 erg/s

0.1 Msol/yr of dust

assuming dust/gas ratio of Long Period Variables

>107Msol over 108 yr outburst lifetime

Metallicity super-solar even in z=6 BELs• High Z/Zsol should enhance

dust production• Larger dust masses likely

Looking Out: quasars & starburstsLooking Out: quasars & starburstsElvis, King et al., in preparation

Conventionally, starbursts fuel quasar outbursts What if it is the other way around?


All Quasars have winds

Quasar wind outflow rates ~1 Msol/yr at L~1046 erg/s

shocks on host galaxy ISM

induces starburst

Fuels AGN

Wind

… cycle of AGN/starburst activity?

Quasar Atmospheres, Quasar WindsQuasar Atmospheres, Quasar WindsGood Observational Paradigm:

Quasar Atmospheres are dynamic

Thin, rotating, funnel-shaped disk wind

Prospects:Use quasar atmospheres for accretion disk physics

Dust creation at high zQuasar to Starburst causality

Low Theory beginnings:

2 phase medium

Line driven winds


Postscript: Imaging QuasarsPostscript: Imaging Quasars

At low z sizes are ~0.1 mas

Resolvable with planned ground interferometers

VLT-I, Ohana

Ideal telescopes:

•Image the wind in space and velocity

•5 km-10 km IR 2m interferometer at ‘Dome C’ in Antartica

•½-1km UV space interferometer

= NASA ‘Stellar Imager’

Quasar community should push for “Quasi-Stellar Imager”

Sizes are implicit in:

Peterson et al. 1999 ApJL 520, 659.

Kaspi et al. 2001 ApJ 533, 631

What we really want is to look at quasar atmospheres

SOLVE QUASAR ATMOSPHERES

No more fancy indirect deductions!

Elvis & Karovska, 2002 ApJ, 581, L67


Solving Quasars part I

Documents