New INTEGRAL High Mass X-ray Binaries Lara Sidoli (INAF-IASF, Milano) Simbol-X Bologna, 2007, May

● A biased introduction to the INTEGRAL discoveries in the HMXB field: highly obscured Galactic sources and... ●....The mistery of the Supergiant Fast X-ray Transients ● IGR J : a unique SFXT ! ● A possible solution to the SFXTs puzzle Overview

from Bodaghee et al., 2007, astro-ph/ Sources detected with ISGRI during the first 4 years of observations

from Bodaghee et al., 2007, astro-ph/ Sources detected with ISGRI during the first 4 years of observations

● X-ray pulsations (slow pulsators, with periods ~ few hundred seconds) ● Hard and highly absorbed X-ray spectra sometimes with strong Fe emission lines (Ph.Index ~0.5-1; absorbing column densities N H > 1E23 cm-2) – the first (discovered on 2003, Jan 29) of these heavily absorbed sources was IGR J (Walter et al 2003, Matt & Guainazzi 2003, Ibarra et al. 2006) ● Optical Identification with Blue Supergiants or Be stars (e.g. Masetti et al. 2005, 2006, Reig et al. 2005) ● They are mainly located in the direction of the spiral arms of our galaxy About 30% of the new IGRs are HMXRBs

The prototype of the highly obscured IGR sources: IGR J IGR J with XMM-Newton & INTEGRAL from Ibarra et al. 2006

Simulation of a IGR J like spectrum with Simbol-X Simulated spectrum (50 ks) Nh = 2E24 cm-2 cutoffpl : ph. index = 1.6 Ec = 60 keV 3 gaussians at 6.4 keV, keV and 7.45 keV

An updated summary of IGR HMXBs with known Pspin or Porb, or both :

Thanks to the Galactic Plane monitoring with INTEGRAL: heavily obscured sources & transients with short outburst have been discovered in the first 4 years of observations

The story of the SFXTs begins with XTE J an unusual new x-ray transient Smith et al. 1998: RossiXTE obs Aug 12 – discovery of a new transient, active only 1 day ! ~ 3E-9 erg/cm2/s (2-25 keV): the brightest source in the direction of the Galactic Center region on that one day Thermal bremsstrahlung spectrum kT ~ 21 keV => Be/XRB ???

XTE J an unusual new x-ray transient Smith et al. 1998: RossiXTE obs Aug 12 – discovery of a new transient, active only 1 day ! ~ 3E-9 erg/cm2/s (2-25 keV): the brightest source in the direction of the Galactic Center region on that one day Thermal bremsstrahlung spectrum kT ~ 21 keV => Be/XRB ??? Negueruela et al. 2005: Optical counterpart is an O-type supergiant (O8 I ) -> dist ~ 2 kpc => a new class of transient sources!?!?!

INTEGRAL discoveries and observations Supergiant Fast X-ray Transients unusually SHORT outbursts, significantly shorter than Be/NS binaries ● OUTBURST DURATION: ~ few hours, less than a day ● RECURRENT OUTBURSTS ● HARD AND HIGHLY ABSORBED X-RAY SPECTRUM – reminiscent of an X-ray pulsar ● OPTICAL COUNTERPART: highly reddened O B supergiants ISGRI SWs (~2 ks) image sequence (20-30 keV) flux at peak=254 mCrab from Sguera et al IGR J / XTE J

A growing number of observations and a growing number of sources mainly thanks to INTEGRAL ! SFXTs and SFXT candidates initially selected because of the short outbursts, then possibly confirmed based on the optical counterpart from Negueruela et al Sguera et al. 2005, 2006 Source Dist Sp.Type Peak flux (20-40) (kpc) ISGRI (erg/cm2/s) IGR J /XTE J O8Iab 6.5 E- 09 IGR J O9Ib 3.0 E- 09 SAX J ? supergiant earlier than B3 1.3 E- 09 IGR J ? ? 2MASS source 2.4 E- 10 IGR J /AX J B0I 9.0 E- 10 IGR J (=SBM10) ? 10 ? E- 09 IGR J /AXJ ? E- 10 AX J ? E- 10 XTE J ? E- 10 IGR J ? - supergiant?? 6.0 E- 10 IGRJ /AX J O 9.5 I 6.0E-10 IGR J ? Ob5Ib 6.0E-09 IGR J Other possible candidates: IGR J , IGR J , IGR J , IGR J , IGR J , IGR J , IGR J ,IGR J

A growing number of observations and a growing number of sources mainly thanks to INTEGRAL ! SFXTs and SFXT candidates initially selected because of the short outbursts, then possibly confirmed based on the optical counterpart from Negueruela et al Sguera et al. 2005, 2006 et al. Source Dist Sp.Type Peak flux (20-40) (kpc) ISGRI (erg/cm2/s) IGR J /XTE J O8Iab 6.5 E- 09 IGR J O9Ib 3.0 E- 09 SAX J ? supergiant earlier than B3 1.3 E- 09 IGR J ? ? 2MASS source 2.4 E- 10 IGR J /AX J B0I 9.0 E- 10 IGR J (=SBM10) ? 10 ? E- 09 IGR J /AXJ ? E- 10 AX J ? E- 10 XTE J ? E- 10 IGR J ? - supergiant?? 6.0 E- 10 IGRJ /AX J O 9.5 I 6.0E-10 IGR J ? Ob5Ib 6.0E-09 IGR J Other possible candidates: IGR J , IGR J , IGR J , IGR J , IGR J , IGR J , IGR J ,IGR J Hard X-ray spectra with high energy cut off at ~ keV L outburst / L quiesc. ~ 1E3-1E4 (from ASCA archival, or XMM or Chandra obs) but only from few sources!

1997, Aug, XTE obs 2004, Aug, ISGRI obs Flux (20-40 keV) at the peak of the outburst, together with the “quiescent” level quiescent level measured well far away from the outbursts

IGR J in't Zand 2005 Chandra obs

SFXTs properties: absorbing column density from X-ray spectra N H – N H Gal (1E22 cm-2) Total N H Galactic (1E22 cm-2) towards the sources

SFXTs properties: X-ray luminosity during the short flares

HMXRB classes: a summary Persistent wind fed accreting X-ray pulsars with supergiant companions “typical” HMXRBs, Vela X-1 like Be/X-ray binaries typically Transient X-ray sources with Be companions A new class -> Supergiant Fast X-ray Transients with supergiant companions transient emission, with short duration outbursts, typically few hours, less than Be/XRBs outbursts highly absorbed HMXRBs discovered with INTEGRAL a growing number of members have been discovered with INTEGRAL Bright persistent disk-fed massive X-ray binaries (Cen X-3 like) in close orbits

Supergiant Fast X-ray Transients: proposed interpretations (I) ● Outbursts produced by short ejections from the donor stars in XRBs? or clumpy winds? (in 't Zand 2005) this suggestion is based on optical observations of the clumping nature of the wind from early type stars: ● Owocki et al. 1997, Lamers & Cassinelli 1999: line-driven winds from supergiant stars are highly variable

Supergiant Fast X-ray Transients: proposed interpretations (II) ● A new kind of supergiant HMXRBs in wide eccentric orbits? (wider than “normal” supergiants HMXRBs, like Vela X-1, in nearly circular orbits) also to explain the large Lmax / Lmin ( Negueruela et al ) Lx expected from an eccentric orbit around a supergiant with a spherical homogeneous wind in the case: 30 Msun companion star beta wind=1 terminal wind velocity=1800 km/s Porb=200 days Wind mass loss rate=2 E-6 Msun/yr

IGR J : a unique SFXT Fast transient discovered on April 22, 2005 (Lubinski et al. 2005) : detected with ISGRI in 2 consecutive pointings, reaching 75 mCrab at peak (20-60 keV) The companion is a B-type Supergiant: HD (Negueruela et al. 2005; Masetti et al. 2006)

ISGRI lightcurves (17-40 keV) 1 st outburst 3-4 July nd outburst May rd outburst 22 Apr 2005 ~ 330 days ! 330 days = likely orbital period The INTEGRAL lightcurve shows a periodicity of ~ 330 days in the recurrence of the outbursts ( Sidoli, Paizis & Mereghetti, 2006, A&A, 450, L9; astro-ph/ ).

4 th outburst lightcurve 4 th outburst lightcurve (Smith et al. 2006, ATel 766 and ATel 773) from XTE/PCA extreme variability X-ray spectrum: hard and quite constant through the outburst Ph. index ~ 1.7 +/- 0.2 ( keV) NH ~ 11 +/- 3 E22 cm-2 (higher than Galactic) Flux at peak ~ 2 x erg/cm2/s roughly consistent with previous outbursts

IGR J : INTEGRAL spectra 1 st outburst 2 nd outburst JEM-X IBIS/ISGRI Energy (keV) Energy (keV) ph.index = ph.index = F = 2.5 E -10 erg/cm2/s ( keV)F = 6.2 E -10 erg/cm2/s (5-100 keV) E cutoff = 15 +/-5 keV L = 3E36 erg/s (5-100 keV)

IGR J is a slow X-ray pulsator in a wide orbit XTE/PCA pointings performed during the 4 th outburst in 2006 showed variability which suggested a pulsation period of approximately 195 +/- 10 s (Smith et al. 2006). The pulsations have been confirmed during the 5 th outburst in Feb 2007 (Swank et al. ATel # 999): Pspin= /- 0.3 s This would imply an X-ray pulsar with: Porb ~ 329 days P spin ~ 190 s thus an accreting pulsar lying in the Be\XRB region of the Corbet diagram ?? BUT the companion is a supergiant! so it is not a Be/NS system !!!  Porb=329 days seem to suggest that outbursts in SFXTs might be related to wide highly eccentric orbits and not to clumpy winds

Monitoring campaign strategy: - 2 ks / day at the beginning (starting on 4 th February) - then exposure up to 10 ks / day during outburst - Monitoring end planned for 14 th February (actually, IGR11215 was kindly observed up to 26 th February) - Thanks to Neil Gehrels and to the Swift Team Results in Romano, Sidoli, Mangano, Mereghetti & Cusumano, 2007 (A&A in press; astro-ph/ ) The periodic nature of the outbursts from IGR J allowed us to plan & perform for the first time a sensitive and complete monitoring of the entire outburst from a SFXT, during the latest outburst, expected on 2007, 9 th February with Swift/XRT

IGR J Final lightcurve from Swift / XRT observations of IGR J th February 2007 Romano et al. 2007

1 st fact: the accretion phase in SFXTs lasts longer than originally thought based on less sensitive instruments

Note that INTEGRAL would have seen ONLY the bright “orange” region of the lightcurve, lasting less than 1 day and composed of several ``short flares” 9 th February 2007

Close-up view of the brightest part of the outburst (9 th Feb) 2 nd fact: a large variability is present, and the outburst event is composed of short flares, lasting few hours or less

N H =0.85E22 ( )  =0.94 ( ) N H =1.11E22 ( )  =0.91 ( ) N H =0.83E22 ( )  =0.82 ( ) N H =0.88E22 ( )  =1.03 ( ) N H =2.02E22 ( )  =1.51 ( )

The total Galactic absorption towards the source is 0.8 E22 cm-2

Model of accretion from a spherical homogeneous wind in an eccentric orbit around the HD star How to explain the outburst? The recurrence time of 329 days is the underlying clock of the phoenomenon, which can be interpreted in a natural way as the orbital period of the binary system

A sorta fairytale 3 rd fact: the lightcurve is too narrow and steep to be explained within a model of Bondi accretion from a spherical and homogeneous wind in an eccentric binary (even with extreme eccentricities!)

Using the neutron star as a probe of the supergiant wind, the shape of the X-ray lightcurve implies that the wind from the B supergiant is NOT spherical nor homogeneous then.....

Using the neutron star as a probe of the supergiant wind, the shape of the X-ray lightcurve implies that the wind from the B supergiant is NOT spherical nor homogeneous then..... our idea is that there is a second wind component in a “thin equatorial disk” with lower wind velocity and a higher wind mass loss rate compared to the “polar wind component”

from Ud-Doula, Townsend & Owocki, 2006, ApJ, 640, L191 Simulation of the radiatively driven outflow from a rotating hot star with a dipole magnetic field Log (wind density)

The proposed geometry to explain the short SFXT outbursts: The thin equatorial disk of the B-supergiant is inclined with respect to the orbital plane The star has also a polar wind with higher velocity and lower mass loss rate (~0.01 * Mass loss rate in the disk) thus, in order to explain the low X-ray emission level out of the outburst, we need anyway a not circular orbit ns OB supergiant Sidoli et al., 2007, in preparation

The thickness “h” of the densest region of this disk h orbital plane “h” can be estimated from the duration of the outburst: the duration t of the brightest part of the outburst is t ~ 1 day, the ns velocity near periatron is roughly vns ~ km/s thus: h ~ 8E11 – 1.7 E12 cm ( Ropt, if Ropt ~ 40 Rsun)

The thickness “h” of the densest region of the disk h orbital plane “h” can be estimated from the duration of the outburst: the duration t of the brightest part of the outburst is t ~ 1 day, the ns velocity near periatron is roughly vns ~ km/s thus: h ~ 8E11 – 1.7 E12 cm ( Ropt, if Ropt ~ 40 Rsun) Nh galactic < 0.8 E 22 cm-2; Nh from X-ray spectra ~1-2 * 1E22 cm-2 Assuming about 1E22 cm-2 is local and mainly due to the supergiant disk, a rough estimate of the   density of the supergiant disk is: Nh*mH / h =  cm 

“disk” wind “polar” wind variable wind density along the orbit

Example of the model of wind accretion with the following parameters: B-supergiant mass = 39 Msun “polar wind” mass loss rate = 3.7 E-6 Msun/yr terminal velocity (“polar wind”) = 1800 km/s beta law exponent = 1.0 Porb = 329 days e = 0.4 Note that along the orbit, the centrifugal barrier is always open (with B ~ 1E12 G, and Pspin=187 s) “disk wind” with Mdot_wind = 100 * Mdot_polar terminal velocity = 900 km/s “disk wind” with Mdot_wind = 30 * Mdot_polar terminal velocity = 1600 km/s

few further questions....

how to explain other SFXTs where a clear periodicity in the outbursts has not been found yet? probably a different eccentricity & a different geometry of the thin disk with respect to the orbital plane

how to explain other SFXTs where a clear periodicity in the outbursts has not been found yet? probably a different eccentricity & a different geometry of the thin disk with respect to the orbital plane how to explain the persistent Vela X-1-like HMXBs with supergiant donors? probably the “equatorial disk” of the supergiant in the persistent systems lies on the orbital plane and the ns always moves inside the disk

Conclusions ● INTEGRAL observations opened a new view on High Mass X-ray Binaries ● The new class of SFXTs has been discovered ● The SFXT IGRJ is a key system to understand this new class of X-ray sources ● We proposed a new hypothesis to explain the properties of the SFXTs, where the supergiant has an equatorial component of the wind, besides the “standard” “polar” high velocity wind. ● The enhanced accretion rate when the neutron star cross this disk produces the short duration X-ray outburst