University of California, Santa Barbara Accretion Disks in AGNs Omer Blaes University of California, Santa Barbara

Collaborators Spectral Models: Shane Davis, Ivan Hubeny Numerical Simulations: Shigenobu Hirose, Neal Turner Simulation Analysis and Theory: Julian Krolik

AGNSPEC -Hubeny & Hubeny 1997, 1998; Hubeny et al. (2000, 2001)

The Good: Models account for relativistic disk structure and relativistic Doppler shifts, gravitational redshifts, and light bending in a Kerr spacetime. Models include a detailed non-LTE treatment of abundant elements. Models include continuum opacities due to bound-free and free-free transitions, as well as Comptonization. (No lines at this stage, though.)

The Bad --- Ad Hoc Assumptions: Stationary, with no torque inner boundary condition. RPtot with  constant with radius - determines surface density. Vertical structure at each radius depends only on height and is symmetric about midplane. Vertical distribution of dissipation per unit mass assumed constant. Heat is transported radiatively (and not, say, by bulk motions, e.g. convection). Disk is supported vertically against tidal field of black hole by gas and radiation pressure only.

LMC X-3 in the thermal dominant state BeppoSAX RXTE Red = BHSPEC, blue=compton component from “corona” (COMPTT), green=total (WABS*(BHSPEC+COMPTT)), and black is data -Davis, Done, & Blaes (2005) The same sort of accretion disk modeling that has been attempted for AGN works pretty well for black hole X-ray binaries (BHSPEC, Davis et al. 2005, Davis & Hubeny 2006).

Some Recent Observational Developments That Have Direct Bearing on Our Understanding Of Accretion Disks in AGN Spectropolarimetry has succeeded in removing BLR, NLR, and dust emission in the optical and infrared, revealing the underlying broadband continuum shape for the first time (Kishimoto’s talk later in this session). Ton 202 Model fit is a face-on disk around a Kerr (a/M=0.998) hole of mass 8.e9 Msun, accreting at 2 Msun/yr. The model is the total flux from the disk - the polarized flux from this disk shows a much stronger edge and is in the wrong direction. -Kishimoto et al. (2004)

(2) Microlensing observations have now placed constraints on the physical size of the optical continuum emitting region in many QSO’s. 0.33 0.1 Disk expectations (the points) give half-light radii a factor 3-30 too small according to their estimates of the minimum half-light radii necessary to attenuate the microlensing variability observed. Paul schechter tells me that there are 2 differences between Pooley et al. and Kochanek: a factor of 3 in the lensing Statistics, and a factor of 5 in the black hole mass Measurements. -Pooley et al. (2006)

-Dai et al. (2006)

(3) Reverberation mapping leveraged by BLR radius/continuum luminosity correlations has given a method of getting approximate black hole masses for the huge number of SDSS quasars. 5100/4000 4000/2200 All models have a/M=0.998 and random cos i>0.5. Their Tmax is entirely dependent on the FWHM of the MgII, and completely independent of the luminosity. They find, however, that Tmax is highly correlated with L/Ledd in their sample, and that L/Ledd=0.3 at Tmax=5.4. Slim disk effects could therefore redden the 4000/2200 color at higher inferred Tmax by reducing the radiative efficiency of the inner disk. 2200/1350 -Bonning et al. (2006)

5100/1350 -Bonning et al. (2006)

AGNSPEC Blackbodies -Davis et al. (2006)

SDSS data AGNSPEC AGNSPEC With E(B-V)=0.04 -Davis et al. (2006) (4000-2200) (2200-1450) AGNSPEC AGNSPEC With E(B-V)=0.04 AGNSPEC models shown here are all Schwarzschild. (Higher spin - 0.9 - is too blue in both colors, and would require more reddening to match.) M and L are Monte-Carlo’ed to give a similar distribution as seen in the data. Luminosity is monochromatic luminosity at 2200 angstroms rest. Masses are measured using monochromatic luminosity at 3000 angstroms and MgII FWHM. The monte carlo assumes a uniform distribution of cos I>0.5, and log Normal distributions in L/Ledd and M_BH, with the mean of M_BH increasing with redshift. Data only include FWHM>~2000 kms, and this is accounted for in Monte Carlo by rejecting models in the wrong part of the parameter space. Shane agrees with this bottom line conclusion: In the absence of internal reddening, the theoretical disk colors are more or less right in the near UV, but are too blue in the UV. A little bit of reddening in the UV solves the latter, and might be marginally consistent with the former (but maybe not!). -Davis et al. (2006)

 begone!!! Thermodynamically consistent, radiation MHD simulations of MRI turbulence in vertically stratified shearing boxes are telling us a lot about the likely vertical structure of accretion disks. Turner (2004): prad>>pgas Hirose et al. (2006): prad<<pgas Krolik et al. (2006): prad~pgas

Radiation Magnetic times 10 Gas 6.62 solar mass black hole, L/Ledd=0.14 (10% efficiency), 150GM/c^2. The t=90 epoch barely satisfies the Shakura-Sunyaev thermal instability criterion (Prad>0.6Ptot): Prad/Ptot=0.64. Average cooling time in disk is ~7 orbits. Gas

Expect strong (but marginally stable) thermal fluctuations at Dissipation rate is proportional to ~E at low energy. Dissipation rate is proporational to ~E^1/2 at high energy. Cooling rate for Erad~Egas is proportiona to E^-1. Hence marginal stability at low energy and stability at high. When Erad>>Egas, cooling rate will probably depend more weakly on E, which might give rise to thermal instability depending on how dissipation scales with E. Expect strong (but marginally stable) thermal fluctuations at low energy and stable (damped) fluctuations at high energy.

Gravity Total Magnetic Radiation Gas

With magnetic fields No magnetic fields CVI K-edge Solid line is with magnetic pressure support, dashed line is without. Both use the same broken-power law dissipation profile that comes from the simulation. This is actually nu Inu, in ergs/cm2/s/sr, viewed at inclination of 55 degrees. -Blaes et al. (2006)

Complex Structure of Surface Layers Photosphere Photon Bubble Shock Train??? Parker

Spectral Consequences Magnetically supported upper layers decrease density at effective photosphere, resulting in increased ionization and a hardening of the spectrum. Strong (up to factor 100) irregular density inhomogeneities exist well beneath photosphere of horizontally averaged structure. They will soften the spectrum. Actual photosphere is therefore complex and irregular, which will reduce intrinsic polarization of emerging photons (Coleman & Shields 1990). Magnetic fields may also Faraday depolarize the photons (Gnedin & Silant’ev 1978):

Overall Vertical Structure of Disk with Prad~Pgas MRI - the source of accretion power Photosphere Parker Unstable Regions Pmag>Prad~Pgas Prad~Pgas>Pmag State that the dissipation is still concentrated toward the midplane, and main vertical heat transport mechanism is photon diffusion, not Poynting flux or bulk fluid motions. Finish talk by saying that we are Monte Carlo’ing the emergent spectrum from this annulus, and a Prad>Pgas simulation is running right now! Pmag>Prad~Pgas