Review of X-ray variability in black hole binaries · Review of X-ray variability in black hole...

Review of X-ray variability in black hole binaries

Adam Ingram Michiel van der Klis, Chris Done 1234

1234

1234

1234

1234

1234

The Extremes of Black Hole Accretion 9th June 2015

102 5 20 50

0.1

110

Flux

Energy (keV)

102 5 20 50

0.1

110

Flux

Energy (keV)

102 5 20 50

0.1

110

Flux

Energy (keV)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

0.01 0.1 1 10

10−4

10−3

0.01

(rms/

mea

n)2

Frequency (Hz)

Wijnands & van der Klis (1998)

Truncated disk model

1 10 1000.01

0.1

110

Flux

Energy (keV)

Flu

x

State changes from moving truncation radius (Rg=GM/c2)

Ro~60Rg

Ro~6Rg

Gilfanov (2010)

4

Diffusion 5

Lynden-Bell & Pringle (1974) Lyubarskii (1997)

Diffusion 6

Lynden-Bell & Pringle (1974) Lyubarskii (1997)

Propagating fluctuations

r1 r2 r3

0.1 1 10

10−4

10−3

Powe

r

Frequency (Hz)

1/tvisc(r1)

1/tvisc(r2)

1/tvisc(r3)

7

Lynden-Bell & Pringle (1974); Lyubarskii (1997); Arevalo & Uttley (2006)

1234

1234

1234

Propagating fluctuations

30 35 40 45 50 55

0.10.2

r1 r2 r3

Accr

etio

n Ra

te a

t r1

Accr

etio

n Ra

te a

t r2

Accr

etio

n Ra

te a

t r3

X-ra

y Flu

x

0.1 1 10

10−4

10−3

Powe

r

Frequency (Hz)

1/tvisc(r1)

1/tvisc(r2)

1/tvisc(r3)

7

Lynden-Bell & Pringle (1974); Lyubarskii (1997); Arevalo & Uttley (2006)

Time (s)

Propagating fluctuations

r1 r2 r3

Propfluc: Ingram & Done (2011, 2012) 0.1 1 10 100

10−3

0.01

Freq

uenc

y x

Powe

r

Frequency (Hz)

8 1234

1234

1234

30 35 40 45 50 55

0.10.2

Accr

etio

n Ra

te a

t r1

Accr

etio

n Ra

te a

t r2

Accr

etio

n Ra

te a

t r3

X-ra

y Flu

x

Time (s)

0.04 0.06 0.08 0.1

0.01

0.02

0.03

RMS

(abs

olut

e)

Flux

Propagating fluctuations

r2 r3

Propfluc: Ingram & Done (2011, 2012)

8 1234

1234

1234

30 35 40 45 50 55

0.10.2

Accr

etio

n Ra

te a

t r1

Accr

etio

n Ra

te a

t r2

Accr

etio

n Ra

te a

t r3

X-ra

y Flu

x

Time (s)

r1

0.1 1 10 100

10−3

0.01

5×10

−42×

10−3

5×10

−30.

02

Freq

uenc

y x P

ower

([σ/µ]

2 )

Frequency (Hz)

2 rings: χν2=1.008

10 rings: χν2=0.980

20 rings: χν2=0.980

30 rings: χν2=1.014

Propagating fluctuations

r1 r2 r3

Propfluc: Ingram & van der Klis (2013)

8 1234

1234

1234

30 35 40 45 50 55

0.10.2

Accr

etio

n Ra

te a

t r1

Accr

etio

n Ra

te a

t r2

Accr

etio

n Ra

te a

t r3

X-ra

y Flu

x

Time (s)

variability has a softer spectrum than fast variability

Freq-resolved spectra 9

Slow and cool

Revnivtsev, Gilfanov & Churazov (1999) Gilfanov, Revnivtsev & Churazov (2000)

Propagation time Fast and

hot

Slow and cool

Kotov et al (2001); Arevalo & Uttley (2006)

Propagation time

Time lags

(8.2-14.1keV) vs (0-3.9keV)

0.1 1 10 100

10−4

10−3

0.01

0.1

Tim

e la

g (s

)

Frequency (Hz)

2 rings: χν2=1.006

10 rings: χν2=0.987

20 rings: χν2=0.996

30 rings: χν2=0.997

Fast and hot

Cyg X-1: Nowak et al (1999) Propfluc: Ingram & van der Klis (2013)

10

Hard photons lag soft

Slow and cool

Kotov et al (2001); Arevalo & Uttley (2006)

Propagation time

Time lags Hard photons lag soft Harder photons: greater lag

Fast and hot

11

Energy (keV) 10

Cyg X-1: Kotov et al (2001)

QPO: Frame dragging A spinning black hole distorts space and time

The satellite’s motion is influenced by the spin of the black hole

Lense & Thirring (1918) Stella & Vietri (1998)

12

Ingram, Done & Fragile (2009)

0 5 10 15 20

4000

6000

X−ray B

rightnes

s

Time (s)

GRS1915+105

QPO: Frame dragging 13

Flux

Energy

Tell-tale sign of precession: a rocking iron line

Ingram & Done (2012)

https://www.youtube.com/watch?v=e1QmLg5mGbU

QPO: Frame dragging 14

Flux

Flux

Time

X-ra

y Br

ight

ness

2045 2050 2055 2060 2065

40006000

8000X−ra

y Brightnes

s

Time (s)

GRS1915+105

Energy Energy

Phase Resolving 15

QPO waveform

0 1 2 3−20

24

68

Total

2nd Harmonic

1st Harmonic

Cycles

16

Periodic function: constant phase difference

0 1 2 3−20

24

68

QPO waveform

Total

2nd Harmonic

1st Harmonic

Cycles

16

Periodic function: constant phase difference

QPO waveform

…but does the phase difference vary randomly or around a well defined mean?

1900 1920 1940 1960

40006000

8000X−ra

y Counts

/s

Time (s)

GRS1915+105

17

Quasi-periodic function: changing phase difference

QPO waveform Quasi-periodic function: changing phase difference

Split long light curve into many segments and measure the phase difference ψ for each segment

1900 1920 1940 1960

40006000

8000X−ra

y Counts

/s

Time (s)

GRS1915+105

17

QPO waveform Phase difference varies around a mean: there is an

underlying waveform

0 0.5 1 1.5 2

0.5

1N

o of s

egm

ents

(nor

mal

ised

)

ψ / π Ingram & van der Klis (2015)

18

GRS 1915+105 (0.46 Hz QPO)

0 0.5 1 1.5 2

4000

5000

6000

Cou

nts/

s

Phase (QPO cycles)

Obs 1

Obs 2

Phase difference varies around a mean: there is an underlying waveform

QPO waveform

Ingram & van der Klis (2015)

19

GRS 1915+105 (0.46 Hz QPO)

Phase resolving Spectra for 4 snapshots of phase

20

105

0.8

11.

2R

atio

Energy (keV)

Obs 1GRS 1915+105 (0.46 Hz QPO)

Ingram & van der Klis (2015)

Spectral modeling

QPO phase-resolved spectroscopy 7

105

0.8

11.

2R

atio

Energy (keV)

Obs 1

105

0.8

11.

2R

atio

Energy (keV)

Obs 2

Figure 8. The spectrum corresponding to 4 QPO phases, plotted as a ratio to the mean spectrum. Phases are selected to be representative of rising (blue), peak(black), falling (red) and trough (green) intervals. For observation 1, these phases are φ = 0, 0.1875, 0.625 and 0.75 QPO cycles respectively. For observation2, they are φ = 0, 0.3125, 0.4375 and 0.625 cycles. For both observations, we clearly see spectral pivoting.

4000450050005500

C/s

Observation 1

2.22.32.42.5

K30.7m

0.80.9

1

kTbb

(keV

)

26.1m

6.26.36.4

E c (ke

V) 1.9m

0.20.40.6

m (k

eV) 0.6m

0 0.5 1100150200250

EW (e

V)

Phase (QPO cycles)

2.4m

3000350040004500

C/s

Observation 2

22.052.1

2.15

K 50.4m

0.20.40.6

kTbb

(keV

)

1.0m

6.16.26.36.4

E c (ke

V) 0.3m

0.40.60.8

1

m (k

eV) 0.9m

0 0.5 1100200300

EW (e

V)

Phase (QPO cycles)

3.6m

Figure 9. Best fit parameters of our spectral fit plotted as a function of QPO phase. Both observations show highly significant spectral pivoting and we see amodulation in the relative strength of the iron line. The statistical significance of each modulation is quoted in the corresponding plot.

Ingram & van der Klis (2015)

21

3

From NuSTAR to PolSTAR

CZT detectors!

boink!

φ"

LiH scattering rod!

Same CZT detectors!

Image!γ"

γ"

photon detected!

(Launch: 6/2012)!

Compton scattering "in LiH rod!

φ"0! 2π!

# photons!Use to "compute"degree of"polarization!

10/30/14 Copyright 2014 California Institute of Technology. U.S. Government sponsorship acknowledged. log10(flux)

Ro=20Rg, a=0.98

www.youtube.com/watch?v=ieZYYfCapJg&feature=youtu.be

3

From NuSTAR to PolSTAR

CZT detectors!

boink!

φ"

LiH scattering rod!

Same CZT detectors!

Image!γ"

γ"

photon detected!

(Launch: 6/2012)!

Compton scattering "in LiH rod!

φ"0! 2π!

# photons!Use to "compute"degree of"polarization!

10/30/14 Copyright 2014 California Institute of Technology. U.S. Government sponsorship acknowledged.

3

From NuSTAR to PolSTAR

CZT detectors!

boink!

φ"

LiH scattering rod!

Same CZT detectors!

Image!γ"

γ"

photon detected!

(Launch: 6/2012)!

Compton scattering "in LiH rod!

φ"0! 2π!

# photons!Use to "compute"degree of"polarization!

10/30/14 Copyright 2014 California Institute of Technology. U.S. Government sponsorship acknowledged.

Ingram et al (2015)

Polarization 22

•  Propagating fluctuations model consistent with power spectrum, linear rms-flux relation, time-lags, frequency-resolved spectra…

•  Can now do propfluc analytically, so fitting lots of data is feasible (see Stefano Rapisarda’s talk)

•  If the QPO is due to precession, the iron line shape should change with QPO phase

•  QPO phase-resolved spectroscopy is now possible (see Abi Stevens’ talk)

•  Need to look at more observations

23

Conclusions