Tadayuki TakahashiInstitute of Space and Astronautical Science (ISAS)

Spectral and Temporal Variations of Blazars

Hidetoshi Kubo(Kyoto), Jun Kataoka (Tokyo IT), Chiharu Tanihata (ISAS)

Cosmic-ray and Particle Accelerator

• Where are accelerators?

• What’s the maximum energy?

• How powerful?• “Black Holes” are

important players?

Blazars are ideal objects tostudy the behavior of particleaccelerators at the bottom of Jets

3C46 (1.4 GHz)

core (AGN) + inner jet

knot

lobe

hot spot

Yes:AGN

Cosmic Ray Spectrum

Gamma-ray Blazars

• Gamma-rays – Direct evidence of the exi

stence of GeV/TeV particles

– the emitting source cannot be

• too compact• too close to important sour

ces of X-ray photon (e.g. a hot accretion disk corona close to the black hole)

to avoid γγ-> e+e-

EGRET sky map of AGNs

TeV detection

Third EGRET AGN Catalogue

Mrk 421

Takahashi et al. 1996,1999Macomb et al. 1995

ASCA

Whipple (TeV)

EGRET

X-Gamma CorrelationsX and γ-rays are cospatial

Takahashi et al. 1999

1995

Fossatti et al. 2003

Gamma-ray Blazarsproduced in relativistic jets pointing close to the line of sight

observerBLK

Enhancement

by Relativistic Beaming(Lobs 〜 L4)BLKcos

　

Dominated by non-thermal highly variable

broad-band radiation

High PowerSmall Emax

Low PowerHigh EmaxKataoka 2002

Fit to Spectral Energy Distribution(SED) -> Parameters of Acceleratorsbased on the assumption of

–Synchrotron–Inverse Compton

•Sync. Photon (and External)•Peak frequency relations•Lumunosity relations

Solve “Parent” electron distributionfrom the spectra

• Self-consistent analysis can constrain– Size– Magnetic Field– Beaming Factor– Electron Distribution

(Kinetic Power)

X-ray band is sensitive to γmax and γmin

max = 105

5x105

min = 1 min = 1000

X-ray

Temporal Variations (TeV Blazars)at the maximum end of electron distribution

Takahashi et al. 1999

time (x 10,000 s)

cnts/s

1day

2000

2000

ASCA’s Long-look Observation (Still Difficult for XMM/Chandra)Daily FlareShape : almost symmetric : Light Crossing Effect in the blob

(not the effect of cooling/acceleration time constant)Offset Component (pedestal)

Spectral Variations (TeV Blazars)at the maximum end of electron distribution

Kataoka et al. 2001

Tanihata et al. 2003

Takahashi et al. 2000

Acceleration/Cooling

High if

shock velocity (vs) is high

　 or is high

€

∝B

€

∝1

B2γ

€

τ acc = τ coolτcool (Obs. Frame)B=0.1 Gauss= 10 0.5 keV … 17,000 s 5 keV … 5,000 s

(at X-rays)

Flare light curve is symmetric.No energy dependence foundin rise/decay.

Time dependent treatment• Time dependent modeling is important to study th

e spectral evolution (but available only very recently)– time scale of

• Acceleration and Cooling• Escape

(Kirk et al. 1999, Kataoka et al. 2000, Krawczynski et al. 2002)

• Predicts characteristic spectral evolution (such as “soft lag/hard lag”) , from the balance between τacc and τcool.

New Component (ex. with higher γmax)

Kataoka 2000

B=0.1 gauss

R=1016 cm3

…

Flare Light Curve

Injection

escape

Solve

the time evolution of electrons

tvar

Characteristic Time Scale

1 day

Mrk421

Mrk501

PKS2155-304

-Daily flares are commonly observed -Characteristic time scale : tvar 〜 40ks - 100ks

-

-Both-Structure Function-PSD

analysis indicate

time variablity <tvar is greatly suppressed for t < tvar

R 〜 ctvar 〜 1016 [cm]

Internal Shocks

€

R ~ ctvarΓ

tvar ~R

cΓ~D

cΓ 2

(R ~ Dsinθ ~ Dθ ~D

Γ)

€

tcatchup ~ 2Γ 2τ

D ~ 2cτΓ 2

€

tvar ~ τ

0.5 - 1 day variabilityroughly corresponds to 10 Rg for 109 M

(Rees 1978Ghisellini 1999, Spada et al. 2001Kataoka et al. 2001)D ~ BLK

2 D0 ～ 1017 [cm]

R ~ BLK D0

~ 1016 [cm]

d ~ 2D0

B.H.

shock

BLK

τ

(Kataoka et al. 2001,Iwashimizu et al. 2003)

Fast shell catches up the slow shell

Link to the Characteristic time scale of the ejection from BH.

Variablity

Light Curve Simulation

- Blobs mainly collide at 　 D ~ 103-4 D0 = 1017-20 [cm]

- ｍ = 10, D0 = 3×1013 [cm]

- Only the flares due to collisions at the smallest distance will be appeared as “shots (daily flares)”

Day-by-dayflares

Internal Energy

Flare time-scale (ksec)

No. of flaresoffset

log D (cm)

Time

Flarge

OFFSETsmaller

structure function

offset component

Tanihata et al. 2003

Simulation Observation

EUVE1keV6.3keV15keV

Rfo =0.7, Tchr =40 ks ⇒ D 0 =1x1013 cm

　　　 G =0.015, =15 (assumed) -> very narrow distribution is required

Light Curve (Flare, Energy dependent Amplitudes, SF/PSD are OK. Efficiency <0.01 %

Application to Mrk 421

Similar Analysisby Guetta et al. 2002provides Efficiency < several %still small

avrg

assumption

Energy carried out by the form of Jetdoes not go into electron acceleration/radiation

One more issue to tackle with

Absorption effect (TeV γ)by Diffuse Extragalactic Background Radiation

F. Aharonian 2003

Need to correct TeV spectra for the SED fitting, if the emission exceeds several TeV

Re-visiting SED AnalysisD

B.H.

shockshock

Takahara et al. (Poster 81)

1. IR absorption corrected for TeV spectra (important)

2. Fit quiescent phase to determine parameters. Use higher for flare, scale other parameters with

3. Collision takes place at longer distance for larger

4. maximum energy is higher for larger

Quiescenceδ=12, B=0.12G

δ=37, B=0.012GFlare in 2000

for Mrk 421Flare

Change of the parameters ofaccelerator ?

Another approach to fit FLARE

Paradigm Shift

Inhomogenious Model

(continuous flow)

Before CGRO/TeV/ASCA

Homogenious One Zone Model

After X-GammaCorrelation

Time-DependentOne Zone Model

After Detections ofFlare & Spectral Evolution

Time Dependent(Internal Shocks)

Multi Zone?

After CharacteristicTime Scale (Daily Flare)

Future Observations

ISAS

GLAST2006

AstroE2 GLAST

Kataoka et al. 1999

GLAST SKY

?

poster by Fukazawa et al.

Conclusion• We have a fairly good understanding of Blazar Spectra

(Parameters of Accelerators); ue>uB

• X-ray/TeV correlations give strong constraint on the model

• Low Efficiency in sub-pc jets (Blazar emission)– Most of the energy carried out from BH is transported to kpc-j

ets and Lobes (See Poster 32 by Kataoka)• Shift of Paradigm

Time Dependent Model is indispensable Internal Shock Model (Multi Zone?)

• Need sensible and Detector in hard-X and Gamma

Narrow FOV Compton Telescope for the NeXT mission in Japan

–Stack Configuration•Low Energy 24 layers of Strip Strip detectors (res. 400μm) and 6 mm thick CdTe Pixel (res. 1mm)

–High Energy Resolution of <1 - 3 keV

BGO

•Incident angle of γ-rays are defined by a well-type active collimator (Extremely Low Background)

ISAS