1. Magneto centrifugal winds from accretion discs around black hole binaries by Susmita Chakravorty from Institut de Planétologie et d'Astrophysique de Grenoble (IPAG) OSUG, Université Joseph Fourier with Pierre-Olivier Petrucci, Jonathan Ferreira, Gilles Henri The Extremes of Black Hole Accretion 9 th June, 2015

2. Winds in black hole X-ray binary Credit: Margaret Masetti (GSFC) Stellar mass black holes are found in binary systems. These systems undergo semi-periodic out bursts and become bright in X-rays. During outburst the observed luminosity performs a “hysteresis” wrt spectral hardness. The binaries pass from Hard to Soft state and back. The X-ray continuum for the two states are very different. 10M ʘ black hole accreting at 0.1 ṁ Edd Thermal – Diskbb; r in = 6 r g  T in = 0.75 keV Powerlaw -  = 2.5 L disk /L PL = 0.8 in 2 – 20 keV Hard – Diskbb; r in = 20 r g  T in = 0.39 keV Powerlaw -  = 1.8 L disk /L PL = 0.2 in 2 – 20 keV

3. Winds in black hole X-ray binary – an overview Neilsen et.al. 2012 FeXXVI FeXXV There are 20 confirmed black hole binaries (Remillar & Mclintock 2006) But 4 BHBs show absorption lines (RXTE + Chandra or XMM-Newton) Most observations show absorption lines from ‘only’ FeXXV and FeXXVI (black spectra) Exceptions (?) GROJ1655, 2006 observation (Miller et.al. 2008) – has numerous lines (blue spectra) GRS1915, 2000 observation (Lee at.al. 2002, Ueda et,al. 2010)

4. Neilsen et.al. 2012 There are 20 confirmed black hole binaries (Remillar & Mclintock 2006) But 4 BHBs show absorption lines (RXTE + Chandra or XMM-Newton) Most observations show absorption lines from ‘only’ FeXXV and FeXXVI (black spectra) Exceptions (?) GROJ1655, 2006 observation (Miller et.al. 2008) – has numerous lines (blue spectra) GRS1915, 2000 observation (Lee at.al. 2002, Ueda et,al. 2010) Ponti et.al 2012 The presence of winds is a “state dependant” effect Winds are observed in the Soft state Further, winds are observed in objects of high inclination (i.e. low equatorial angle) Winds in black hole X-ray binary – an overview

5. Can these solutions represent observable winds (in terms of , N H, n H and v obs ) – both average ones and extreme ones. Can we recover the (i) state dependent and (ii) angle dependant observability? If these efforts are successful, then we shall apply the same methods to AGN winds. Compare the predictions of MHD driven with those from thermally driven models MHD winds from the accretion disk: the ANR-Chaos project f(n) f(v) f(B) f(dyn) ======== Miller et.al. (2008), Kallman 2009 suggest MHD winds from spectra of GROJ 1655 Fukumura 2010a,b, 2014, 2015 – MHD outflows in AGN explaining BALs, WAs, UFOs. (Contopoulos, 1994; Blandford and Payne, 1982) BHB winds have never been modelled with MHD solutions. Begelman (1983), Blandford & Begelman (1999), Proga (2000, 2002), Proga et.al. (2003), Netzer et.al. (2006), Luketic et.al. 2012 etc. have studied thermal winds for BHB accretion disk outflows. (Jonathan) Ferreira, 1997 MHD models are adopted for modeling the wind – providing f(x). Also Casse & Ferreira 2000 and Ferreira & Casse 2004. The ejection or outflow of material is related to the accretion mechanism. Thus, ejection efficiency is **not** a free parameter (unlike ADIOS scenarios) The solutions are self similar. Hence can spread out to large distances. Two relevant fundamental parameters: Disk aspect ratio  (= h/r) Accretion efficiency p (where ṁ acc = r p )

6. The cold MHD outflow  = h/r = 0.001 p = 0.04 ( Ṁ acc = r p )  ~ 1/p

7. Selection of the  constraint Ponti et.al 2012

8. Selection of the  constraint Imposing the constraint N H < 10^24 The detectable wind  = 0.001 p = 0.04

9. Imposing the constraint N H < 10^24 The detectable wind  = 0.001 p = 0.04  = 0.001 p = 0.04 We perform this exercise for a range of  and p We check the distance density velocity of the “closest wind point”

10. p = 0.01  = 0.001 Cold (purely) magnetic solutions do not work (?) The effect of  is always lesser. Hence we concentrate on the effect of variation of p Note that the Hard SED, itself, does not make any significant effect difference from Soft SED! So the intrinsic flow has to be different to explain “ winds in Soft state” The angles of Line of sight agree with Ponti et.al. 2012. Winds can be detected for low equatorial angles (high inclination angles.) And finally – the wind is just too far away, the density and velocity are too low. So is there a way to make things better? We need to increase p. But within cold solutions, no super-Alfvenic winds with larger p. Warm MHD solutions – surface disk heating lifts material. In this framework, p can be increased further. I (degree) log [R sph (cm)]

11.  = 0.01 p = 0.10  = 0.01 p = 0.10 Cold  = 0.001 p = 0.04 Warm magnetic solutions

12. Our Best Warm solution In the given framework, it was difficult to increase p further. We are working on it ….. why? Fukumura et.al. papers - p ~ 0.5! So there is room to improve. A rough linear extrapolation puts the wind at 5x10 4 R G Choice of  upperlimit is a key player in determining the wind region. We had chosen a rather stringent upperlimit. Relaxing it to log  < 6 brings the wind closer by ~ 30 times (actual calculations on  = 0.01 and p = 0.10 solution) at ~ 5 x 10 4 R G at ~ few x 10 3 R G log n H ~ 12.4, log vobs > 4

13. Warm magnetic solutions Our Best Warm solution In the given framework, it was difficult to increase p further. We are working on it ….. why? Fukumura et.al. papers - p ~ 0.5! So there is room to improve. A rough linear extrapolation puts the wind at 5x10 4 R G Choice of  upperlimit is a key player in determining the wind region. We had chosen a rather stringent upperlimit. Relaxing it to log  < 6 brings the wind closer by ~ 30 times (actual calculations on  = 0.01 and p = 0.10 solution) The change in density and velocity would depend on the MHD solution. Extrapolating linearly – they would increase through about 2 orders of magnitude. In fact, the numbers will be more extreme, when we consider the radiative transfer along the line of sight at ~ 5 x 10 4 R G at ~ few x 10 3 R G log n H ~ 12.4, log vobs > 4

14. The MHD models used are more realistic because the accretion and ejection are inter related It is difficult to explain observed winds with cold (purely) MHD solutions The wind is equatorial – consistent with observations. Over-ionization of plasma is not the reason for wind disappearance in Hard state. The underlying outflow has to change. Dense (largest p = 0.11 that we obtain) warm MHD solutions can explain the average BHB winds. - disk surface heating increases the amount of available material -which is then accelerated by magnetic fields For the most spectacular winds (Miller et.al. 2008, King et.al. 2012), we need denser (p  0.5) MHD winds. Can we obtain them? – work in progress. Conclusions  ~ 1/p Ṁ acc = r p

15. Absorption due to LOS material

16.  = 0.01 p = 0.10 Effect of thermodynamic stability Hard Thermal log  = 1 2 3 4 5 Compton heating 79% Compton heating 36%