Accretion flows onto black holes Chris Done University

Accretion flows onto black holes Chris Done University of Durham

Lectures 1 -3 BHB for flow and jet • Dramatic changes in continuum – single object, different days • Underlying pattern in all systems • High L/LEdd: soft spectrum, peaks at k. Tmax often disclike, plus tail • Lower L/LEdd: hard spectrum, peaks at high energies, not like a disc Ljet (Mc. Clintock & Remillard 2006) DGK 07

Lectures 4 -6 AGN • Scale up to AGN! • Same transition • Disc temperature lower (UV not X) • Obscuration – torus and host galaxy!! • Jet can be very different G=1. 5 in BHB and LINERS, 10 -20 in Bl Lacs/FRI and FSRQ/FRII

Unobscured AGN: LINERS-S 1 -NLS 1 Boroson 2002 Similar mass. Different L/LEdd Different ionisation Increasing L/LEdd disc Hot inner flow, no disc – true Seyfert 2 s NLS 1 S 1/QSO LINER

Unobscured AGN: LINERS-S 1 -NLS 1 Boroson 2002 Similar mass. Different L/LEdd Different ionisation Increasing L/LEdd disc Hot inner flow, no disc – true Seyfert 2 s NLS 1 S 1/QSO LINER

Classic QSO – most common shape Medhipour et al 2011 • Mkn 509 - 108 M L/LEdd~0. 1 (take out warm abs!) • Not disc dominated - far too low temperature! Plus strange soft X-ray excess…. What is this? ?

Classic QSO? Medhipour et al 2011 • Mkn 509 - 108 M L/LEdd~0. 1 • Not disc dominated - far too low temperature! Plus strange soft X-ray excess…. What is this? ?

Standard QSO • Spectra generally not dominated by the disc • hard tail and soft X-ray excess carry a large fraction of Lbol • IR from torus reradiation of UV Richards et al 2006, Elvis et al 1994

Unobscured AGN: LINERS-S 1 -NLS 1 Boroson 2002 Similar mass. Different L/LEdd Different ionisation Increasing L/LEdd disc Hot inner flow, no disc – true Seyfert 2 s NLS 1 C I M S O C T I A YS S 1/QSO PH LINER

X-ray absorption: neutral • H edge 13. 6 e. V = 0. 013 ke. V • Higher Z elements have higher edge energy for K shell electron as higher charge means inner electrons more tightly bound • Outer electrons shielded so ionisation energy is less • CNO K 0. 28, 0. 40, 0. 53 ke. V ionisation 9, 11 and 14 e. V • Fe K & L edges at 7. 1 and 0. 7 ke. V n=1 K shell n=2 L shell n=3 M shell…etc

X-ray absorption: neutral H • Higher Z elements less abundant so total absorption cross section decreases with energy Log s He C Log E N O

X-ray absorption: ionisation • Leaves ion! • Ion can recombine if more free electrons than X-ray photons • Then its back to neutral before the next X-ray comes. So Xrays only see neutral material • BUT what if the X-ray comes before the electron. Ion is not neutral and all edge energies are higher as unbalanced charge n=1 K shell n=2 L shell n=3 M shell…etc

X-ray absorption: ionised H • Higher Z elements less abundant so total absorption cross section decreases with energy Log s He C Log E N O

Photoionised absorption: edges • • if completely ionised then no edges left!! Just power law ionised edges are higher energy net charge so more tightly bound H like edge at 0. 0136 Z 2 ke. V (energy charge/r) so high Z elements need more energy to completely ionise. Fe K He, H-like 8. 7, 9. 2 ke. V (XXV and XXVI). if dominant then everything else is ionised! x=103 x=102 Nh=1023 x=1

Photoionisation: populations • which ions? Balance photoionisation (heating) with recombination (cooling) • Depends mostly on ratio of photon to electron density! • ng/Ne = L/(hn 4 p r 2 c Ne) = x / (hn 4 p c) x = L/ (Ner 2) • Nh, x, AND spectral shape Ni + g Ni+1 + e Ni ng s = Ni+1 Nea(T) Ni+1 = ng s Ni Nea(T)

Photoionisation: populations Ni + g Ni+1 + e • Another way to define is ratio of photon pressure to gas pressure • Prad = X = L 1 Pgas 4 pr 2 c nk. T = x / (4 pck. T) Ni ng s = Ni+1 Nea(T) Ni+1 = ng s Ni Nea(T)

Photoionised absorption: edges • • if completely ionised then no edges left!! Just power law ionised edges are higher energy net charge so more tightly bound H like edge at 0. 0136 Z 2 ke. V (energy charge/r) so high Z elements need more energy to completely ionise. Fe K He, H-like 8. 7, 9. 2 ke. V (XXV and XXVI). if dominant then everything else is ionised! x=103 x=102 Nh=1023 x=1

Photoionised absorption: edges • x=103 Multiple edges as generally multiple ion states not just one x=102 Nh=1023 x=1

Lines: even neutral material! • • • K edge energy is 1 s - generally not that much bigger than 1 s-2 p Ka line eg H edge at 13. 6, Lya 10. 2 e. V can see (just) for C N O with good resolution data but EW is generally small compared to edge don’t see this from neutral high Z elements as L shells filled for Z> Ne (Si, S Fe…) but can when ionise! Which also means hotter material 1 -2 Ka 1 -3 Kb 1 -4 Kg 1 - K

Lines: ionised! • • • See 1 s-2 p if got hole in L shell One electron less than filled 2 p shell ie one electron less than neon like Still need at least 1 electron so F-like to H-like has LOTS of lines He like generally biggest crosssection O: 0. 6 ke. V Fe: 6. 7 ke. V 1 -2 Ka 1 -3 Kb 1 -4 Kg 1 - K

Ionised absorption: lines!!! • • BIG difference: LINES absori does ionisation balance and corresponding edge absorption xion does ionisation balance (better as balances heat/cooling) and line + edge absorption Use this where material close enough to X-rays to be ionised!!

Evidence for Winds in AGN: X-ray absorption • See ionised absorption lines in soft X-ray spectra of around 50% of nearby AGN Reynolds et al 1997 • ‘warm absorbers’ ie ionised material • With good grating spectra see its multiphase Blustin et al 2005 – eg NGC 3783 has at least 3 different x • V~500 km/s outflow!!! NGC 3783 Netzer et al 2003

Evidence for Winds in AGN: UV absorption • X absorption also predicts UV absorption. • Major ‘high ionisation’ UV transitions eg CIV are low ionisation X-ray absorbers. • Big absorption from low column as UV opacity >> X-ray opacity • See these, v~500 km/s…. (NAL – narrow absorption lines, some line up directly with X-ray absorbers Mathur et al 1999) MR 2251 -187 Monier et al 2001

How to make a wind? ? • EDDINGTON • Reach LEdd in inner regions, so wind from inner disc • Winds typically have velocity ~v_esc(Rlaunch) • Inner disc is 0. 1 -0. 3 c!! 30, 000 -100, 000 km/s • This is >> 500 km/s • And most AGN L<Ledd • So this can’t be the origin of the warm absorbers

Other radiation driven winds • But any material above disc is also illuminated by the continuum source • Effective gravity is (1 - t/tes L/LEdd) GM/R NOT simply GM/R • If just electron scattering = (1 - L/LEdd) GM/R continuum driven wind • Exceeds Ledd in central regions of disc so wind from inner disc!

2: UV line driven Winds ? Log nfn • If substantial opacity: t>>tes so gravity (1 - t/tes L/LEdd) GM/R • Most opacity in UV resonance lines • Momentum absorbed in line accelerates wind so more momentum absorbed in line - UV line driving at L<<LEdd Log E

2: UV line driven Winds ? Log nfn • If substantial opacity: t>>tes so gravity (1 - t/tes L/LEdd) GM/R • Most opacity in UV resonance lines • Momentum absorbed in line accelerates wind so more momentum absorbed in line - UV line driving at L<<LEdd Log E

2: UV line driven Winds ? • • Surprisingly hard to do as X-ray source as well ! UV bright disc launches UV line driven discwind. Rises up and accelerates away unless the central X-ray/UV overionises it so that no UV transitions! Proga 2003

2: UV line driven Winds ? • • • Surprisingly hard to do as X-ray source as well ! UV bright disc launches UV line driven discwind. Rises up and accelerates away unless the central X-ray/UV overionises it so that no UV transitions! But failed wind region can shield (Murray & Chiang 1998, Risaliti & Elvis 2010) - BAL QSO Proga 2003

Evidence for Winds in AGN: UV/opt absorption • BAL QSOs: broad, smooth blueshifted abs. 0 -0. 15 c in 10% QSOs • Broad permitted lines often absorbed to blue in UV

Evidence for Winds in AGN: UV/opt absorption • BAL QSOs: broad, smooth blueshifted abs. 0 -0. 15 c in 10% QSOs • Broad permitted lines often absorbed to blue in UV

3: UV continuum opacity and dust driven winds ? • gravity (1 - t/tes L/LEdd) • Neutral gas has much bigger cross-section • Dust has even bigger cross-section! • But only survives if T<1600 K • L=As. T 4 !! • Sets the inner edge of the torus • Illuminated torus beyond this point should be outflowing! Fabian et al 2006

Roth et al 2012 3: UV continuum opacity and dust driven winds ?

3: UV continuum opacity and dust driven winds ? • Maybe BLR arises from dust driven wind Czerny & Hryniewicz 2011 • Then sees X-ray source so dust evaporates – dust free BLR • Some evidence from data – Galianni & Horne 2013 show a universal disc temperature at the Hβ radius ≈ 1600 K, close to dust sublimation Czerny & Hryniewicz 2011

4: themal winds - Compton temperature Log nfn • X-ray heating of material from compton up and downscattering De/e=4 Q - e • Integrate over number of photons N(e) at each energy • ∫ N(e) De = 0 = ∫ N(e) (4 Q -e) e de • Compton temperature TIC=511 QIC 4 QIC= ∫ N(e) e 2 de / ∫ N(e) ede Log n

Compton temperature Log nfn • X-ray heating of material from compton up and downscattering De/e=4 Q - e • Integrate over number of photons N(e) at each energy • ∫ N(e) De = 0 = ∫ N(e) (4 Q -e) e de • Compton temperature TIC=511 QIC 4 QIC= ∫ N(e) e 2 de / ∫ N(e) ede Log n

Thermally driven Winds • Direct illumination or scattering from wind… • X-ray source irradiates top of disc, heating it to Compton temperature • TIC depends only on spectrum - Lirr only controls depth of layer Begelman Mc. Kee Shields 1983

Thermally driven Winds • Hot so expands as pressure gradient – corona bound if v 2 =3 k. TIC/m <vesc 2 = GM/R IC • Wind for R > RIC driven by pressure gradient so expands on cs with v∞= (3 k. TIC/mp) = (GM/RIC) • Wind velocity typically that of gravitational potential from where it is launched R=RIC Begelman Mc. Kee Shields 1983

Absorption lines in BHB Kubota et al 2007 Neutral Ionised ISM absorption

Absorption lines in BHB • • He–like Fe 6. 7 ke. V H-like Fe at 7. 0 ke. V Ratio 6. 7/7. 0 gives x Increasing so ionisation state decreasing with L as expect for photoionsed material To get column need width of line < 4000 km/s Guess a ‘reasonable’ number and get Nh x=L/nr 2 = L Dr/(Nh r 2) Assume Dr/r~1 to get r= x Nh /L distance of material from X-ray source

Thermal winds from BLR and torus? R IC

X-ray Warm absorbers and UV NALS in AGN • few hundred km/s typical velocity for narrow X-ray and UV absorbers • Thermal winds from BLR and torus? • Not much kinetic power – not as much as required for the AGN to affect the host galaxy evolution • Blustin et al 2005

Evidence for Winds in AGN: UFOs • 0. 1 -0. 3 c winds, much larger columns, much higher kinetic luminosity – AGN feedback!!! • Highly ionised, only Fe left with bound electrons (Reeves et al 2009) • So can’t be UV line driving – but these objects are very UV bright and X-ray weak which is what you need – geometry maybe? Or MHD? Hagino et al 2014

5: magnetically driven Winds Fukumura et al 2014

Winds in AGN 1. super. Eddington – need L/Ledd > 1 2. UV line driven - need UV Teff>30, 000 K ! And FEW X-rays 3. Dust driven - need dust: Teff<1400 K 4. Thermally driven – need to be far away so k. T(IC) > GMm/R 5. Magnetohydrodynamic – need ordered B field geometry ? ? • • • UFO (>15, 000 km/s) UV and/or MHD BAL (5000 -20, 000 km/s) UV line driving BLR (few thousand – 10, 000 km/s) UV or dust X-ray warm absorbers (few hundred km/s) thermal or MHD UV NAL (few hundred km/s) thermal or MHD NLR – few hundred km/s thermal or MHD

Classic QSO – most common shape Medhipour et al 2011 • Mkn 509 - 108 M L/LEdd~0. 1 (take out warm abs!) • Not disc dominated - far too low temperature! Plus strange soft X-ray excess…. What is this? ?

Classic QSO? Medhipour et al 2011 • Mkn 509 - 108 M L/LEdd~0. 1 • Not disc dominated - far too low temperature! Plus strange soft X-ray excess…. What is this? ?

So what do AGN look like? • More standard BLS 1/QSO <M>~108, <L/LEdd>~0. 1 • Not disc dominated so can’t use to get BH spin Jin et al 2012

Simple NLS 1 • spectra are much dominated by disc at L~LEdd! • NLS 1 have small SX and big disc - NOT huge SX Jin et al 2012

Conclusions • Lots of DIFFERENT winds • Thermal wind from torus - Warm absorber/associated narrow UV absorption lines • Dust wind from disc – BLR? ? • UV line driven disc wind for BALs – CAN work if Xray/FUV weak which we see at high L/Ledd, high M • High L/Ledd>1 also means get continuum wind! • Should be able to quantify AGN wind feedback from UV line driven winds as a function of M and L/Ledd (and spin)

AGN/QSO Zoo!!! Radio loud • Enormous, powerful, relativistic jets on Mpc scales • FRI (fuzzy) - BL lacs FRII (hot spots) – FSRQ • Urry & Padovani 1992; 1995

FRI is top of ADAF branch (low/hard state BHB) but G=15! L/LEdd BHB Ghisellini et al 2010

Optically thin nonthermal compton • power law by single scattering of nonthermal electrons N (g) g-p • index a = (p-1)/2 (p >2 so a > 0. 5 – monoenergetic injection) • Starts a factor t down from seed photons, extends to gmax 2 ein eout~g 2 ein f(e) e-a e-(p-1)/2 Log fn Log N(g) g N (g) g-p Log g gmax ein Log n gmax 2 ein

Nonthermal synchrotron • power law by single scattering of nonthermal electrons N (g) g-p • index a = (p-1)/2 (p >2 so a > 0. 5 – monoenergetic injection) • Starts a factor t down from seed photons, extends to gmax 2 ein eout~g 2 ein f(e) e-a e-(p-1)/2 Log fn Log N(g) g N (g) g-p Log g gmax ein Log n gmax 2 ein

Synchrotron self compton Log vfn • Put in vfv • Expect index a = (p-1)/2~0. 6 for p=2. 2 from shocks gmax ein Log n gmax 2 ein

Synchrotron self compton Log vfn • Klein nisina cutoff – can’t have more energy than electron had to start with g 2 hn < g mc 2 so g e < 1 where e=hn/mc 2 • Synchrotron self absorption gmax ein Log n gmax 2 ein

Synchrotron self compton Log vfn • Klein nisina cutoff – can’t have more energy than electron had to start with g 2 hn < g mc 2 so g e < 1 where e=hn/mc 2 • Synchrotron self absorption gmax ein Log n gmax 2 ein

Synchrotron self compton Log vfn • Doppler boosting due to bulk motion G (don’t confuse with lorentz factor of electrons g) – d = 1/[G(1 -bcosq)] • Eobs=Eint d • Fobs=Fint d 3+a gmax ein Log n gmax 2 ein

Synchrotron self compton Log vfn • What we see in BL Lacs (Tavecchio et al 2010)

BL Lacs as SSC • need G ~ 10 -20, q<5 o, gmax~105 double power law electron distribution • Can’t make FSRQ Ghisellini et al 2010

FSRQ –disc and BLR • Disc is behind jet so strongly deboosted • BLR may be much more isotropic so these external seed photons can be more important self produced synchrotron Ghisellini et al 2009

FSRQ –disc and BLR Log vfn • Disc is behing jet so strongly deboosted • BLR may be much more isotropic so these external seed photons can be more important self produced synchrotron gmax ein Log n gmax 2 ein Ghisellini et al 2009

FSRQ –disc and BLR Log vfn • need G ~ 10 -20, q<5 o, gmax~105 double power law electron distribution similar to BL Lacs gmax ein Log n gmax 2 ein Ghisellini et al 2009

Kinetic luminosity of the jet • Estimate Pjet from models (only small fraction of power is radiated) • See disc as well as jet so compare Pjet with Pdisc • About equal in FSRQ

Kinetic luminosity of the jet • Can’t do this in BL Lacs so easily • but many FRI are in clusters of galaxies • Jets expand as bubbles in hot cluster gas – so measure total energy input from Pd. V work

Conclusions • Should be able to quantify AGN jet feedback as a function of M and L/Ledd (and spin) • UV line driven disc winds depend on spectral shape – this is observed to depend on L/Ledd. Weak hard X-rays and strong UV predict more powerful winds – see this at high L/Ledd and high mass. • We can constrain relativisitic jet power observationally – but we see these jets predominantly from most massive black holes. What triggers the relativisitic jet ? ?