X-ray Variability of AGNOutlineTiming VariabilityPower Spectral Density (PSD)Slide 5What does the PSD tell us?Long/Short Term VariabilityPSD of NGC 4051Slide 9The AGN-BHXRB Connection“Low” & “High” StatesNGC 4051 & Cyg X-1Slide 13Are AGN & BHXRB similar enough to use one to study the other?Physical InterpretationsModels to Account for PSDsRMS VariabilityTime LagsOptical VariabilityOptical/X-ray CorrelationsCause of Optical/X-ray VariabilityNot Simple AbsorptionSpectral VariabilityContinuum VariabilityFe Kα (Line) VariabilityFractional Variability Amplitudes (Fvar)Measured FvarModels/MechanismsVariations to Simple ModelMore Variations to the Simple ModelBroadband Spectral VariabilityConclusionsReferencesX-ray Variability of AGNJudith RacusinAstro 597AWeek 5October 6, 20042 of 33Outline•Timing Variability–PSD–RMS–GBH/BHXRB analogy to AGN–Time lags in UV/Optical•Spectral Variability–Fvar–Lines–Continuum–Models•Conclusions3 of 33Timing Variability•A defining characteristic of AGN•Time scale of months to years, suggests scale of emitting phenomena (one of the first clues of AGN BH connection)•No obvious periodic behavior in light curves•Variable behavior on large range of time scales•McHardy et al. (2004) present observations using RXTE and XMM-Newton of AGN (NGC 4051) & GBH(Cyg X-1 ) comparing various properties of these object mostly related to their PSDs4 of 33Power Spectral Density (PSD)•Variability can be characterized by PSD functions to understand the behavior •“PSD defines the amount of variability ‘power’ as a function of temporal frequency” 1•PSD ~ |Discrete Fourier Transform|2•P  f -α5 of 33Vaughan et al. (2003)6 of 33What does the PSD tell us?•Break timescale–Scales with mass–Breaks from index of ~2 to ~1–Needed to avoid divergence•Amplitude of PSD in ~day-long light curves  1/Lx–Lx  MBH–tvar  BH size–This works only for BLS1 not NLS1Uttley et al. (2004)7 of 33Long/Short Term Variability•AGN with different luminosities are very different on short time-scales, but similar on long time-scales•Length of observations may not spread across all long-term variability•AGN appear to be “red noise” dominated (i.e. vary more on long time scales)8 of 33PSD of NGC 4051Fit by unbroken power law over frequency range 7x10-9 to 2x10-5 Hz has slope of -1.059 of 3310 of 33The AGN-BHXRB Connection•Power law shaped PSDs are similar to the “high” state PSDs of black hole X-ray binary systems (BHXRB)•If BH mass scales linearly with break timescale, AGN should have break timescales of days-weeks•This all assumes that the same phenomena is indeed acting on both kinds of systems11 of 33“Low” & “High” States•“high” state => high luminosity/soft x-ray spectrum•“low” state => low luminosity/hard x-ray spectrum•Distinguished by second break in low state PSD where slope = 0 about a decade below the high-frequency break•High state seems to match some AGN well, but low state does not12 of 33NGC 4051 & Cyg X-1•NGC 4051 has best studied AGN PSD•Cyg X-1 is a well studied BHXRB•NGC 4051 PSD has same shape as Cyg X-1 PSD in the high state, scaled to higher frequency.PSD has slope of ~ -1 at low frequencies and steepens to ~ -213 of 3314 of 33Are AGN & BHXRB similar enough to use one to study the other?•Similar PSDs that seem to scale with mass•Evidence from the rms-flux relation observed in BH and neutron star XRBs suggest that variability originates in the accretion flow itself (i.e. not caused by coronal flares)•Note (kind of obvious now): Assuming we believe that BHXRBs have BHs in them, this provides indirect proof that AGN have BHs. This is because they both show PSD breaks where as XRBs (neutron stars) do not but GBH (galactic black hole) candidates do.15 of 33Physical Interpretations•X-ray emission is produced by Comptonisation of low energy seed photons by energetic electrons in a corona•The break timescale may correspond to the viscous or thermal time scales of ~few RG•The non-flat slope of the PSD above the break suggests that the corona in not of uniform temperature, or fluctuations within the corona are responsible for the variability at high frequencies•Additional parameters needed to explain variance between high and low state PSDs: perhaps accretion rate or BH spin (either would change edge of accretion region depending on relation to last stable orbit)16 of 33Models to Account for PSDs•Varying the temperature or density of the emitting region associates different photon energies with different radial locations–Accretion flow propagates inwards in an optically thin corona over the surface of the disc until it hits the X-ray emitting region–Higher energies and shorter timescales occur at smaller radii–Characteristic timescale perhaps relates to viscous timescale at edge of X-ray emitting region–Period dependant lags (lower frequencies propagate slower)17 of 33RMS Variability•RMS variability is measured for small segments of X-ray light curves in a specific band•Mean and variance of segment calculated and then rms = sqrt(variance) in each bin•Both AGN & BHXRBs show strong linear correlation between RMS amplitude of variability and X-ray flux•Conclusion: more variance when object is brighter18 of 33Time Lags•Time-scale dependent lags between hard and soft X-ray•Optical/UV emission suffers from slight lag to X-ray emission – suggests same mechanism as X-ray emission but reprocessed in the disk•Lags are greater for longer Fourier period and increase with the energy separation between bands•Lags are also asymmetric towards positive lags indicating the presence of complex delays of higher energy band lightcurves compared to lower energy•Therefore, higher energies and higher frequencies are associated with smaller radii19 of 33Optical Variability•Long time scales require long, well-sampled monitoring campaigns•Optical variability also red noise•Smaller amplitude of variability than in X-ray•Reverberation mapping from varying of optical emission lines in response to continuum variations–Maps line emitting regions–Gives mass estimates•Continuum emission thought to be primarily thermal from the accretion disk20 of 33Optical/X-ray Correlations•Some AGN show strong X-ray/Optical correlations, others do not•Optical flux is at times larger than X-ray flux, ruling out optical source being pure