1. Accretion Disk Spectra of Ultra- luminous X-ray Sources and Galactic superluminal jet sources Ken Ebisawa (INTEGRAL Science Data Center, NASA/GSFC) Piotr Zycki (Copernicus Center), Aya Kubota (ISAS), T. Mizuno (Stanford), K. Watarai (Kyoto) ApJ, Nov 10 issue, 2003

2. Introduction and Summary “Too-hot” disk problem in ULXs and Galactic super- luminal jet sources –Observed disk temperature is too high for Standard Schwarzschild disk Standard Kerr disk spectra can be much harder when disk is highly inclined –Reasonable for near edge-on Galactic super-luminal jet sources –May not explain hard spectra of ULX (inclination random) Slim disk can explain L>L Edd, high disk temperature, and spectral variation of ULXs –ULXs are black holes with a few tens M  with super-Eddington luminosities

3. Ultra-luminous X-ray Sources (ULX) Discovered with Einstein in nearby spiral Galaxies (e.g., Fabbiano 1988) L X (0.5-10 keV)~10 39 -10 40 erg s -1 Too bright for X-ray binaries, too dim for AGN Most sources are located off-center of the Galaxy (Colbert and Mushotzky 1999) >100 M  not to exceed the Eddington limits?

4. Characteristics of ULX Significant time variation (Source1 in IC342; Okada et al. 1998) Compact object in nature Most probably black holes

5. Characteristics of ULX High-low transition? (Source1 and 2 in IC342; Kubota et al. 2001) Orbital modulation? (Sugiho et al. 2001, Bauer et al. 2001) Similar to Galactic black hole candidates

6. “Too-hot disk” problem in ULX and superluminal jet sources ULX energy spectra –Thermal spectrum, like standard optically thick accretion disk –Disk temperature too high for given luminosity and mass, assuming Schwarzschild black hole (Okada et al. 1998; Makishima et al. 2000) Same problem in Galactic superluminal jet sources GRS1915+105 and GRO J1655-40 (Zhang, Cui and Chen 1997)

7. Standard Schwarzschild disk model fit gives too large luminosity or too small mass Observed accretion disks are “too hot” compared to what is expected from reasonable mass and luminosity Schwarzschild disk best-fit M=1.8 M  L= 0.4 L Edd M=9.4 M  L= 11 L Edd M=100 M  L= L Edd M= 7 M  (real mass) L= 0.1 L Edd

8. Makishima et al. (2000) M=const, L disk ~T in 4 Super-Eddington OK with standard model Too small mass Too large mass accretion rate ULXs GRO J1655 and GRS1915

9. How to explain the “too-hot” accretion disk? Standard accretion disk around Kerr black hole might explain the hard disk spectra (Zhang, Cui and Chen 1997; Makishima et al. 2000) –R in = 3 R s (Schwarzschild)  0.5 R s (extreme Kerr) –Higher disk temperature possible than the Schwarzschild case

10.  In the Schwarzschild metric, R in = 3 R s = 6 GM/c 2  Max disk color temperature for Schwarzschild black hole T col ~ 1.3 keV ((T col /T eff )/1.7) (M/M Edd ) 1/4 (M/7M  ) -1/4  In the Kerr disk, inner radius smaller, temperature can be higher.. Standard accretion disk temperature

11. Laor, Netzer and Piran (1990) extreme Kerr disk Scwarzschild disk Inclined Kerr disk is brighter in high energies! cos i

12. Relativistic effects and inclination dependence of the disk spectra When the disk is face-on, relativistic effects do not change disk spectra significantly Edge-on Kerr disk has very hard spectrum –Due to doppler boosts in the innermost region Kerr disk Schwarzschild disk Newtonian disk

13. Application of Kerr disk spectra to GRO J1655 GRO J1655-40 –i=70 , d=3.2 kpc, T col /T eff =1.7 fixed –M=16 M  with a=0.998 (extremely Kerr) –M=7 M  suggests a=0.68 to 0.88 (Gierlinski et al. 2001) –Inclined Kerr disk model works to solve the “too-small mass problem”! –450 Hz QPO (Strohmayer 2001) supports presence of the disk around a spinning black hole (Abramowicz and Kluzniak 2001)

14. Application of Kerr disk spectra to ULX IC342 Source 1 –face-on Kerr disk (d=4Mpc, T col /T eff =1.7,a=0.998) M=29 M  and L=14 L Edd Super-Eddington problem not solved –edge-on (i= 80  ) Kerr disk (a=0.998) M= 355 M  and L=0.9 L Edd Super-Eddington problem may be solved, but unreasonably large mass required –Standard Kerr disk model not plausible for ULX disk inclination should be random ~350 M  black hole unlikely

15. Adjusting inclination angle does not give reasonable mass and luminosity Standard disk model cannot explain the ULX energy spectra Fitting IC342 spectrum with different inclination angle

16. Slim disk Optically thin Optically thick Abramowicz et al. (1995) Standard disk Soft state Advection Dominated Accretion Flow (ADAF) Hard state Slim disk Optically thick ADAF disk (slim disk) is stable unstable

17. Luminosity of slim disk –Optically thick and geometrically thick disk For standard disk h/r~0.1, L disk < L Edd For slim disk h/r ~ 1  Slim disk luminosity can be ~ 10 L Edd ~10

18. Temperature profile （ Watarai et al.2000 ） Standard disk Characteristics of the Slim disk Radial dependence of the disk temperature becomes weaker ｐ = 0.75  0.50 Advection  Emissivity in the inner disk is suppressed Slim disk (optically thick ADAF) Mup.. Slim disk Standard disk （ radiation dominate) T increase

19. Observation of the slim disk ● Fit with p-free disk XTE J1550-56 Tin p  2 /dof p-free MCD p=0.75 Ldisk saturate Multicolor disk blackbody with T(r)~ r ‐ p p-free disk model Slim disk is observed in Galactic BHC when luminosity is extremely high! Smaller p gives systematically better fits! Kubota 2001 PhD thesis

20. Slim disk in GRS1915+105 TimeEnergy PCA light curves Energy spectra Soft component oscillation Yamaoka 2001 PhD thesis R in T in Standard disk model fit (T(r)  r -0.75 ) High T in T(r)  r -0.5 favored  slim disk Honma, Kato and Matsumoto (1991) Disk oscillation between the standard disk and slim disk

21. Slim disk for ULXs Watarai et al. (2001) L disk saturates at high T in (due to advection) IC342 spectral change explained well slim disk standard disk (L disk ∝ T in 4 )

22. Slim disk model for ULX Fitting ASCA IC342 Source 1 spectrum with Watrai’s slim disk model (face-on, T col /T eff = 1.7, pseudo-Newtonian potential) –M=23 M , L disk ~ 6 L Edd –Super Eddington luminosity allowed with slim disk –Slim disk model fit successful with reasonable mass and disk luminosity!

23. Spectral variation of IC342 Source1 Standard disk fit Slim disk fit Mass varies –Unreasonable, bad model! Consistent mass, only mass accretion rate change –Reasonable, better model!

24. Origin of ULXs Universal luminosity function of X- ray binaries (Grimm et al. 2003) Luminosity cut-off ~10 40 erg/s Presumably, ULXs are X-ray binaries with possible maximum stellar-size black holes ~40 M  black holes theoretically possible (Fryer 1999) Such a massive black holes do not exist in our Galaxy, but not uncommon in external galaxies ULX

25. Summary “Too-hot” disk problem in ULXs and Galactic super- luminal jet sources –Observed disk temperature is too high for Standard Schwarzschild disk Standard Kerr disk spectra can explain the hard spectra of Galactic super-luminal jet sources –Galactic super-luminal jet sources are edge-on systems Slim disk can explain L>L Edd, high disk temperature, and spectral variation of ULXs ULXs are black holes with a few tens M  with super- Eddington luminosities