High X-ray Spectroscopy of V404 Cygni in Near Eddington Outburst

Ashley L. KingStanford/KIPAC

John Raymond, Jon M. Miller, Mark Reynolds,Warren Morningstar

Einstein Symposium Oct. 28, 2015

THANKS!!!!!

• Thank you Belinda Wilkes and the Chandra team for observing this ToO, V404 Cyg during this extremely bright outburst!!!

• Thank you Herman Marshall, David Huenemoerder, and the Chandra Calibration team for your Help and Advice!!!

X-ray Binary Outflows• Markoff • Corbel et al. • Belloni et al.

May1989 Outburst• Zycki et al. 1999

• Ginga Observations

• Strong X-ray variability

• both intrinsic and absorption

• Not clear how long the outburst lasted ~ a few days to weeks

!"#$"%&# '( )* !"!#! ##$ +,-.- /012 %&#

'() *)+,-)* ')./)+0'1+) ,2 34 234' " "!!5 6)7 08* '() 83+.09:

,;0',38 34 '() <3./38)8' <3++)2/38*2 '3 '() ).,'',8= 0+)0 34

! # # 5"& 6.!! $>"5=%!? @A'+0/390')* '3 93B)+ )8)+=,)2C '(,2

<3./38)8' '()8 <38'+,D1')2 2,=8,4,<08'9E '3 '() '3'09 )8)+=E

D1*=)'C B,'( 0 '3'09 D393.)'+,< 491A 34 5"!& )+= <.!! 2!5? F(12

'() 93B)2' /322,D9) 91.,832,'E ,2 0+318* '() @**,8='38 9,.,'C D1' ,'

,2 9,6)9E '3 D) +0'()+ (,=()+ 2,8<) '() 234' 2/)<'+1. .12' D)

D+30*)+ '(08 0 2,8=9) D90<6D3*E 08* '() 0D23+/',38 .0E D) +0'()+

(,=()+ '(08 '() G/33+9E 683B8H ,8')+2')990+ -091) 34 % # 5"!5 <.!!?

!"# $%&'()*+&' ,-./)-%*0&)*-% .-12'3

I <1':344 /3B)+ 90B ,2 83' 8)<)220+,9E 0 =33* 0//+3A,.0',38 '3 0

J3./'38,;)* 2/)<'+1.? K) 4,+2' 12) '() !"#$%& .3*)9 02 08

0809E',< *)2<+,/',38 34 '() 2/)<'+1.C ,8<91*,8= ,'2 2)94:<382,2')8'9E

<09<190')* +)/+3<)22)* 2/)<'+1.? K) 0923 ,8<91*) 0 2)/0+0') L)

M! 9,8) 4+3. 4913+)2<)8<) 4+3. *,2'08' .0')+,09? F() /+)2)8<) 34

'(,2 B02 ,84)++)* DE N32')+D+3)6 )' 09? G5>>&H 4+3. ',.,8= 0809E2,2

34 '() O0E #" *0'0? P,8<) ,' ,2 4+3. *,2'08' .0')+,09 B) 4,A ,'2

)8)+=E 0' &?Q 6)7C 08* 0221.) '(0' ,' ,2 80++3B? K) 0923 ,8<91*) 0

80++3B M" 9,8) 0' R?"% 6)7C ',)* '3 55 /)+ <)8' 34 '() ,8')82,'E 34

'() 80++3B M! 9,8)? F() D)2':4,'',8= !"#$%& .3*)9 (02 #! "#R!#"!R *34C B,'( 93B )9)<'+38 ')./)+0'1+)C 34) " >!!! "!Q 6)7C+0'()+ 90+=) 3/',<09 *)/'(C $)2 " &!R#! "!"RC D1' +)S1,+)2 '(0' '()2))* /(3'38 D90<6D3*E ')./)+0'1+) ,2 -)+E .1<( (,=()+ '(08 '(0'

34 '() 3D2)+-)* 234' 491AC B,'( 4" " "!$#! "!"Q 6)7? L,A,8= '()

2))* /(3'382 '3 '() ')./)+0'1+) 34 '() 3D2)+-)* 234' )A<)22 =,-)2 0

-)+E .1<( /33+)+ 4,' 02 ,' ,2 180D9) '3 +)/+3*1<) '() <1+-0'1+) 2))8

,8 '() (0+* 2/)<'+1. 0' 93B )8)+=,)2? F(,2 D)2':4,'',8= .3*)9 0923

+)S1,+)2 0 2'+38=9E ,38,;)* 08* 2.)0+)* +)/+3<)22)* <3./38)8'C

B,'( 0./9,'1*) %+ ! "!5# G.3*)9 T ,8 F0D9) 5H?

I 2,=8,4,<08'9E D)'')+ 4,' <08 D) 3D'0,8)* 0221.,8= 0 D+30*)+

2/)<'+09 *,2'+,D1',38 34 '() 2))* /(3'382 43+ J3./'38,;0',38? K)

(0-) <382'+1<')* 0 -)+2,38 34 '() !"#$%& .3*)9 ,8 B(,<( '()

,8/1' 234' /(3'38 2/)<'+1. ,2 0 *,2< D90<6D3*EC +0'()+ '(08 0

2,./9) D90<6D3*E? F(,2 .3*)9 G.3*)9 J ,8 F0D9) 5HC B,'( ,'2

<3++)2/38*,8= +)/+3<)22)* <3./38)8'C <08 83B 0*)S10')9E

*)2<+,D) '() *0'0C B,'( D)2' 4,' #! " !%!R"!R *34 G,?)? &#! "55!& <3./0+)* '3 /+)-,312 <02)H? F() J3./'38,;,8= <931* ,2

+0'()+ 2,.,90+ '3 '(0' *)+,-)* /+)-,3129E? U' ,2 3/',<099E '(,<6C

$)2 " &!Q>! "!"RC 08* (02 ')./)+0'1+) 34 34) " >!&""!R!"!% 6)7? F()

0./9,'1*) 34 '() +)/+3<)22)* <3./38)8' ,2 .1<( 2.099)+ '(08 5C

%+ " "!5R""!5&!"!"%C D1' '() +)49)<'3+ /+3/)+',)2 0+) 8)-)+'()9)22 B)99

<382'+0,8)*? U' (02 '3 D) 2'+38=9E ,38,;)* G' ! &"""C 23 V:9,6) ,+38

,382 *3.,80')H 08* 2'+38=9E 2.)0+)*C 5,8 " &5=? F() 80++3B

W0122,08 0' &?Q 6)7 ,2 2',99 +)S1,+)*C 08* ,'2 )S1,-09)8' B,*'( ,2

"R" )7? F() D)2':4,'',8= .3*)9 2/)<'+1. ,2 /93'')* ,8 L,=? #?

I8 09.32' )S1099E =33* 4,' $#! " !>"!R% <08 D) 3D'0,8)* DE

+)/90<,8= '() W0122,08 9,8)2 DE 08 182.)0+)*C 8)1'+09 +)49)<')*

2/)<'+1.C 23 ,' ,2 83' /322,D9) '3 *,2',8=1,2( 4+3. '() 2/)<'+1.

B()'()+ '() *,2'08' .0')+,09 ,84)++)* 4+3. '() ',.,8= 0809E2,2

GN32')+D+3)6 )' 09? 5>>&H ,2 3/',<099E '(,<6 3+ '(,8 43+ )9)<'+38

2<0'')+,8=?

I (0+* 93B)+ 9,.,' 38 6VC )S109 '3 '() 9,6)9E ,8')+2')990+ -091)

34 % # 5"!5 <.!!C B02 ,./32)* ,8 '() 0D3-) 4,'2? I993B,8= 6V '3

D) 0 4+)) /0+0.)')+ ,8 .3*)9 J B) 0<'1099E 4,8* '(0' '() D)2':

4,'',8= -091) ,2 "C 09'(31=( '() 4,' ,2 83B D)'')+ DE 389E &#! " !!&G7:')2' 2,=8,4,<08<) ! >" /)+ <)8'H? F(,2 .0E ,8*,<0') 0 /322,D9)

<3./9)A,'E 34 '() 234' <3./38)8'? L3+ )A0./9)C 0**,8= 0

D90<6D3*E <3./38)8' '3 .3*)9 J B) 3D'0,8 2,.,90+ ,./+3-).)8'

34 '() 4,'C &#! " #!!C B,'( +0'()+ 93B ')./)+0'1+) 34 '()

D90<6D3*EC 34 ! "!5 6)7? U' ,2 83' /322,D9) '3 <(0+0<')+,;) '()

234' <3./38)8' .3+) /+)<,2)9EC 2,8<) '() 93B)+ 9,.,' '3 )891:

D08*/022 ,2 !5 6)7C D1'C ,./3+'08'9EC '() /0+0.)')+2 34 8),'()+

! 5>>> XIPC OYXIP 456C %&5Z%R%

7*89:2 #" J3931+Z<3931+ /93' 43+ *0'0 3D'0,8)* 38 O0E #"? F,.) D,8 ,2 5 2?

F() <,+<9) .0+62 '() +)=,38 4+3. B()+) *0'0 43+ '() 180D23+D)* 2/)<'+1.

B)+) )A'+0<')* GP)<',38 QHC B(,9) '() +)<'08=9) .0+62 0 2,.,90+ +)=,38 43+

'() 0D23+D)* 2/)<'+1. 2(3B8 ,8 L,=? R? J38'31+2 90D)99)* 8! Q 08* 8! %

.0+6 0//+3A,.0') +)=,382 4+3. B()+) '() 234' 2'0') *0'0 B)+) )A'+0<')*

G<4? L,=? 5[ P)<',38 &H? 7)+E +0/,* 08* <(03',< -0+,0D,9,'E B02 3D2)+-)*

*1+,8= ',.) ,8')+-09 8! RC .32' 9,6)9E *1) '3 /(3'3)9)<'+,< 0D23+/',38

GP)<',38 %H? L3+ <3./0+,238C '() 'B3 D,= *3'2 2(3B '() /32,',38 34 Y3-0

O12<0) 5>>5 38 \0810+E 55 08* \0810+E 5& G2)) P)<',38 Q?QC L,=? %H? F()

<+322)2 2(3B 'E/,<09 )++3+ D0+2 38 '() <3931+2] '() D,==)+ 38) ,2 43+ 8! RC

'() 2.099)+ 38) 43+ )0+9,)+ 3D2)+-0',382?

7*89:2 ;" )891: 9,=(' <1+-) G<318' +0') ,8 5Z#" 6)7[ !:2 ',.) D,82H 38 O0E

#"C B()8 '() 231+<) 491A +)0<()* ,'2 .0A,.1. 08* 0//0+)8'9E 20'1+0')* 02

B)99 02 2(3B)* *+0.0',< -0+,0D,9,'E? F() <318' +0') B02 <3++)<')* 43+

D0<6=+318*C *)0*:',.) 08* 02/)<'? T)=,88,8= 34 =33* *0'0C #"C ,2 Q]#5]5#?

^0'0 43+ '() 180D23+D)* 2/)<'+1. GP)<',38 Q[ L,=? #[ F0D9) 5HC 08* 43+ '()

0D23+D)* 38) 2(3B8 ,8 L,=? R B)+) )A'+0<')* 4+3. ',.) ,8')+-09 8! 5? P/)<'+0

4+3. ',.) ,8')+-092 8! Q 08* 8! % 0+) 0809E2)* ,8 P)<',38 &? F()E 0+)

<382,2')8' B,'( 234'_(,=( 2'0') 2/)<'+0 34 P`FC B()8 <3++)<')* 43+ 0D23+/',38?

Into Quiescence• Gallo et al. 2014

• V404 1989 Outburst into Quiescence

• Falls on “Fundamental Plane of Black hole Activity”

• Radio emission has an exponential dependence on X-ray emission, even at low X-ray luminosity

BH X-ray binary radio/X-ray domain 3

2 OBSERVATIONS AND DATA REDUCTION

2.1 Karl G. Jansky Very Large Array

In order to confirm the marginal-significance detection of Miller-Jones et al. (2011), we required a significantly lower noise level.We observed XTE J1118+480 with the Karl G. Jansky VLA on2013 June 27 and 28 (MJD 56471 and 56472), for 7.5 h on eachday (0830-1600 LST). We observed in full polarization mode withtwo overlapping 1024-MHz basebands, centered at frequencies of4.8 and 5.8 GHz. Each baseband was composed of eight contiguous128-MHz sub-bands, each comprising sixty-four 2 MHz spectralchannels. The array was in the relatively compact C configuration,giving an angular resolution of ⇠ 4 arcsec. A power outage at theVLA site caused us to lose 2 h of observation on June 28th, leavingus a total of 11.3 h of time on source. We offset the pointing posi-tion by 7 arcsec (2 synthesised beams) from the known VLBI po-sition of XTE J1118+480 (Mirabel et al. 2001), to prevent artifactsgenerated at the phase center by correlation errors from creating aspurious source.

3C 286 was used as both bandpass and amplitude calibrator,and the secondary calibrator was J1126+4516, a 0.4-Jy source lo-cated 3.2� away from the target, XTE J1118+480. We reduced andimaged the data with the Common Astronomy Software Appli-cation (CASA) v4.1.0, using standard procedures. The data wereHanning smoothed, and then edited to remove radio frequency in-terference. Bandpass calibration was carried out before the ampli-tude and phase gains were derived for both calibrator sources, usingthe ‘Perley-Butler 2010’ coefficients within the CASA task SETJYto set the amplitude scale (Perley & Butler 2013). The complex gainsolutions derived for the secondary calibrator were interpolated tothe target source, before averaging the resulting calibrated data bya factor of three in frequency to reduce the data volume. Imag-ing and self-calibration were then performed separately for eachday’s observations. The data were imaged out to 15 arcmin, wellbeyond the distance to the half-power point of the primary beam.We used Briggs weighting (robust=1) as a compromise betweensensitivity and suppression of side-lobes from bright sources else-where in the field. We used the w-projection algorithm to preventphase errors due to sky curvature from affecting our deconvolu-tion, and modelled the sky frequency dependence using two Tay-lor terms. The brightest confusing source in the field was NVSSJ111820+475659, with a flux density (without primary beam cor-rection) of 4.3 mJy beam�1. There was sufficient emission in thefield to perform self-calibration, initially solving only for phases,and then for amplitude and phase, down to solution intervals of1 and 5 minutes, respectively. For the shorter solution intervals,data from all spectral windows were combined prior to solving forthe time-dependent gains, to provide sufficient signal-to-noise togive robust solutions. Finally, we combined the two self-calibrateddata sets to provide the deepest possible image (shown in Figure1), reaching an rms noise level of 1.45µJy beam�1. A 3.2� peakwas detected 0.59 arcsec from the predicted source position (tak-ing into account the expected proper motion measured by Mirabelet al. 2001), well within the beamsize of 4.2⇥ 4.0 arcsec2. To im-prove the significance of the detection, we then combined our newdata from 2013 June with archival data from 2010 November, pro-viding an extra 2.4 h of time on source. This reduced the noiselevel to 1.36µJy beam�1 in the region around the target. Fittingthe emission at the target position with a point source in the im-

●●●●●●●●●●●

●

●●●●●●●

●

30 32 34 36 38 40

26

28

30

32

log(LX erg s−1)

log(

L rer

gs−

1 )

XTEJ1118+480GX339−4V404 Cyg

●

Figure 2. Linear regression analysis for V404 Cyg (open green triangles;data from Corbel, Kording & Kaaret 2008), GX339-4 (open orange dia-monds; data from Corbel et al. 2013) and XTE J1118+480 (filled blue cir-cles). For the latter, data are taken from this work (lowest LX and Lr lu-minosity point) plus Hynes et al. (2000) and Fender et al. (2001) for the2000 outburst, and Brocksopp et al. (2010) plus Dunn et al. (2010) for the2005 outburst, which is highlighted by open blue diamond symbols encir-cling the filled blue circles. See Table 1 for a complete list of the best-fittingparameters.

age plane gave a flux density of 4.79 ± 1.45µJy beam�1, wherethe quoted uncertainty represents the rms noise added in quadra-ture to the uncertainty on the point source fit. Approximating theintegrated radio luminosity as the monochromatic luminosity mul-tiplied by the observing frequency, this corresponds to a radio lu-minosity Lr= 8.3 ⇥ 10

25 erg s�1. This assumes a flat radio spec-trum and a minimum synchrotron emitting frequency much smallerthan the observing frequency, though the former assumption is vi-olated by some hard state black hole X-ray binaries; e.g., if a spec-tral index a = +0.5 is assumed (where the flux density scales asS⌫ / ⌫

+a), the resulting integrated radio luminosity is factor about2 lower.

2.2 Chandra X-ray Telescope

XTE J1118+480 was observed with Chandra ACIS-S on 2013 June27 (PI: Gallo; ObsId 14630); the data were telemetered in VeryFaint (VF) mode, with a high-energy cutoff at 13 keV, and analyzedwith the Chandra Interactive Analysis of Observations (CIAO)software, v4.5. Event files were reprocessed with the CIAO scriptCHANDRA REPRO. No flares were detected in the background lightcurve, yielding a net exposure time of 58 ks. The analysis describedbelow was carried out between 0.5-7 keV, where the instrument isbest calibrated.

An X-ray source was clearly detected at a position consistent

c� 0000 RAS, MNRAS 000, 000–000

V404 Cyg

June 2015 OutburstSwift/BAT

June 2015 Outbursthttp://deneb.astro.warwick.ac.uk/phsaap/v404cyg/data/

Chandra HETG Spectroscopy of V404 Cyg7

Energy (keV)

0.1

1.0

10.0

keV

(P

ho

ton

s cm

-2 s

-1 k

eV-1

)

2 5 10

Observation 1

(a)

Energy (keV)

0.1

1.0

10.0

keV

(P

ho

ton

s cm

-2 s

-1 k

eV-1

)

2 5 10

Observation 2

(b)

Fig. 1.— This figure shows the unfolded spectrum of the first (a) and second (b) observations. Emission features are observed in Mg XII,Si XIII, Si XIV, S XV, S XVI, Fe K↵ & K�, Fe XXV, and Fe XXVI. In addition changes in spectral shape and a broad excess at 6.4 keVare noticeable between the two observations.

7

Energy (keV)

0.1

1.0

10.0

keV

(P

ho

ton

s cm

-2 s

-1 k

eV-1

)

2 5 10

Observation 1

(a)

Energy (keV)

0.1

1.0

10.0

keV

(P

ho

ton

s cm

-2 s

-1 k

eV-1

)

2 5 10

Observation 2

(b)

Fig. 1.— This figure shows the unfolded spectrum of the first (a) and second (b) observations. Emission features are observed in Mg XII,Si XIII, Si XIV, S XV, S XVI, Fe K↵ & K�, Fe XXV, and Fe XXVI. In addition changes in spectral shape and a broad excess at 6.4 keVare noticeable between the two observations.

June 22, 2015

June 23, 2015

21 ksec

25 ksec

<F> = 9.5x10-9 ergs/s/cm3

<F> = 1x10-8 ergs/s/cm3

<L> = 6.5x1036 ergs/s

<L> = 8.9x1036 ergs/s

Observation 2

8

(a) (b) (c) (d)

(e) (f) (g) (h)

Fig. 2.— These panels show the data to model ratio of each 3.2 ksec epoch in the first (a–d) and second (e–h) observations. Timeproceeds from bottom to top. The dashed lines are Lyman-↵, dotted lines are He-like triplets, and the dot-dashed lines are the Fe K↵ andK� lines.

MgSi S

Fe K

time

He-like Triplet• i/f ~ 1

• if collisionally excited n~3x1013 cm-3

• If photoexcitation from strong UV field is important

• r= 4x1011 cm

• r/(f +i) ~ 0.5 is intermediate between collisional and photoionized cases

• Could be Photoionized, but with T > Teq = 7x105 K or

• Some of r intensity is P Cygni emission -> supported by the blue-shifts and larger line widths

• Outer Disk - orbit of binary separation is r~2x1012 cm

8

(a) (b) (c) (d)

(e) (f) (g) (h)

Fig. 2.— These panels show the data to model ratio of each 3.2 ksec epoch in the first (a–d) and second (e–h) observations. Timeproceeds from bottom to top. The dashed lines are Lyman-↵, dotted lines are He-like triplets, and the dot-dashed lines are the Fe K↵ andK� lines.

NH = n r ~ 3x1022 cm-2

EM = Emission Measure = n2V = 4πn2r3 ~ 2x1058 cm-3

ξ = L/nr2 ~ 1000

r = nr2/nr = L/ξnr = n2r3/(nr)2 = EM/2πNH2

r ~ 3x1012 cm ~ binary separation -> outer diskn ~ 3x1010 cm-3

Emitting Region Density and Size

P-Cygni Profiles• Highest Fluxes

• >0.1 L_Edd

• Absorption increases to >4000 km/s in the highest ionization lines

• Earlier in the outburst Optical P-Cygni Profiles were detected with velocities of ~4000 km/s

9

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.001

0.010

0.100

1.000

NF

e (p

ho

ton

s cm

-2 s

-1)

obs1obs2obs1obs2

(a)

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.01

0.10

1.00

10.00

EW

(k

eV)

obs1obs2obs1obs2

(b)

1036 1037 1038

L(2-10 keV) (ergs s-1)

-2

-1

0

1

2obs1obs2obs1obs2

(c)

Fig. 3.— Panel (a) shows the line intensity of the narrow Fe K↵ line to the total flux measured between 2–10 keV. A tentative positivetrend is noted between these quantities. Panel (b) shows a negative correlation between the equivalent width of the narrow Fe K↵ and thetotal flux. The extremely high value of 1 keV in the lower flux states suggest that we are not observing the direct continuum, but the outerdisk may be obscuring our view. The resulting continuum is a combination of scattered and reflected light from the outer and inner disk,respectively. Panel (c) shows the phenomenological power-law spectral index versus the total flux. � < 1.4 can not be produced by a simpleComptonization model, and therefore absorption, scattering and reflection from the disk are likely altering our view of the continuum.

(a) (b)

Fig. 4.— Panel (a) corresponds to the 8th epoch in the second observation, and shows the ratio of the data to a phenomological power-lawcomponent. Clear absorption features are observed in all but the Fe K↵ and Fe K� lines. The absorption line width increases from MgXII to Fe XXVI. For contrast, panel (b) shows the same ions from the 6th epoch of the second observation. Only emission is detected inthis epoch.

9

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.001

0.010

0.100

1.000

NF

e (p

hoto

ns

cm-2

s-1

)

obs1obs2obs1obs2

(a)

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.01

0.10

1.00

10.00

EW

(keV

)

obs1obs2obs1obs2

(b)

1036 1037 1038

L(2-10 keV) (ergs s-1)

-2

-1

0

1

2obs1obs2obs1obs2

(c)

Fig. 3.— Panel (a) shows the line intensity of the narrow Fe K↵ line to the total flux measured between 2–10 keV. A tentative positivetrend is noted between these quantities. Panel (b) shows a negative correlation between the equivalent width of the narrow Fe K↵ and thetotal flux. The extremely high value of 1 keV in the lower flux states suggest that we are not observing the direct continuum, but the outerdisk may be obscuring our view. The resulting continuum is a combination of scattered and reflected light from the outer and inner disk,respectively. Panel (c) shows the phenomenological power-law spectral index versus the total flux. � < 1.4 can not be produced by a simpleComptonization model, and therefore absorption, scattering and reflection from the disk are likely altering our view of the continuum.

(a) (b)

Fig. 4.— Panel (a) corresponds to the 8th epoch in the second observation, and shows the ratio of the data to a phenomological power-lawcomponent. Clear absorption features are observed in all but the Fe K↵ and Fe K� lines. The absorption line width increases from MgXII to Fe XXVI. For contrast, panel (b) shows the same ions from the 6th epoch of the second observation. Only emission is detected inthis epoch.

Dat

a/M

odel

Morrison et al. - astrobites

P-Cygni Geometry

Depends strongly on inclination, Mdot, velocity law

Face On

Edge OnHigh Mdot

Low Mdot

Disk Wind Profiles

• Mauche & Raymond

Wind Profiles

Dat

a/M

odel

Black Hole Disk Winds 13

FIG. 1.— The first-order spectra of H 1743−322, 4U 1630−472, GRS 1915+105, and GRO J1655−40 are each shown as a ratio to the best-fit continuum model,initially ignoring the Fe K band. Each model included disk blackbody and power-law components, with power-law indices constrained via broad-band fits tosimultaneous RXTE data, where possible. The He-like Fe XXV and H-like Fe XXVI absorption lines in each spectrum show non-Gaussian structure, indicatingcontributions from related lines and/or multiple velocity components. Evidence of weak emission redward of the absorption lines is also apparent, especially inGRO J1655−40 and GRS 1915+105, and suggestive of disk-like P-Cygni profiles.

Dat

a/M

odel

8

(a) (b) (c) (d)

(e) (f) (g) (h)

Fig. 2.— These panels show the data to model ratio of each 3.2 ksec epoch in the first (a–d) and second (e–h) observations. Timeproceeds from bottom to top. The dashed lines are Lyman-↵, dotted lines are He-like triplets, and the dot-dashed lines are the Fe K↵ andK� lines.

“Typical” X-ray Binaries V404 Cyg

8

(a) (b) (c) (d)

(e) (f) (g) (h)

Fig. 2.— These panels show the data to model ratio of each 3.2 ksec epoch in the first (a–d) and second (e–h) observations. Timeproceeds from bottom to top. The dashed lines are Lyman-↵, dotted lines are He-like triplets, and the dot-dashed lines are the Fe K↵ andK� lines.

Miller et al. 2015

Ponti et al. 2012

Typical X-ray Binary Winds

V404 Cyg

• Covering fraction is HUGE, compared to other X-ray binaries • V_wind = 4000 km/s, xi= 3 (Si/S ratios)

• • = 3x1020 g/s = 5x10-6 Msolar/year

• = 2x1037 ergs/s

• = 0.06 30

V404 Cygni Wind Parameters

⇠ =Lion

nr2

Mw = ⌦µL

⇠Vw

Lw =1

2Mwv

2w

Lw

/Lbol

=⌦µv3

w

2⇠Mw/Macc =

Preliminary X-Star models• Good fit for ~low energy

ions

• Super Solar Abundances

• n=10^10 cm^-3

• resonant scattering into our line of sight

• UV emission

• Two Components

• First

• log xi = 1.8

• N_H = 1.3x10^21 cm^-2

• z_abs = -0.002

• Second

• log xi= 4.35

• N_H = 3x10^20 cm^-2

• z_abs=-0.004

Conclusions• V404 Cyg underwent a outburst in June of 2015 that likely reached and/or exceeded its

Eddington limit

• Strong Variability

• Variable X-ray Emission and Absorption lines

• Emission lines are located in the outer disk (1012

cm) with a density of ~1010

cm-3

• P-Cygni Profiles are observed at the highest flux flares

• Dense but Fast wind (>4000 km/s)

• Mass Outflow rate exceeds the Mass Accretion Rate observed in the X-rays

• Future Work

• Accelerated Disk Wind Models • Physical Modeling of the Spectral Evolution within each epoch • Broad Fe K alpha Line Evolution

• Distance - 2.39+/-0.14 kpc - Parallax

• Mass: 9+0.2

-0.6 M_solar ( 8-12 M_solar)

• Black Hole

• LEdd = 1.1x1039

ergs/s

• inclination: 67+3

-1 degrees

• Donor Star - F0-F5 (low mass)

• Roche-lobe overflow

• Super Solar Abundances (1.7xFe_solar)

• Orbital Period 6.4714 days

• orbital separation r~2x1012

cm

V404 Cygni

Casares & Charles 1994, Casares & Jonker 2014, Sanwal et al. 1996, Khargharia et al. 2010, Miller-Jones et al. 2009

1994MNRAS.271L...5C

A parallax distance to V404 Cyg 3

Fig. 2.— Parallax signature of V404 Cyg in right ascensionand declination. The best-fit position and proper motion havebeen subtracted from the measured astrometric positions. Thegrey points show the displacement of the check source from itsweighted mean position (dotted line), shifted by 0.7mas for clar-ity, and scaled by a factor 1/7 to correct for its relative distancefrom the phase reference source. This illustrates the scale of thesystematic errors.

(2009), confirmed the validity of these results, giving me-dian values differing by less than 1σ from those found bythe SVD fit, albeit with slightly larger error bars owingto the data stripping. Our fitted parallax corresponds toa source distance of 2.39 ± 0.14 kpc.

3. DISCUSSION

3.1. Extinction-based distances

Our fitted source distance is significantly closerthan the best previous distance estimate of 4.0+2.0

−1.2 kpc(Jonker & Nelemans 2004). However, that estimate wasderived using an assumed interstellar extinction of AV =3.3, and did not take into account uncertainties in theextinction. Hynes et al. (2009) fitted the full multiwave-length spectral energy distribution with a template spec-trum for a K0 IV star and found a formal best fit red-dening of AV = 4.0. Indeed, we find that increasing AVto close to 4.0 brings the estimate of Jonker & Nelemans(2004) down into agreement with our results. This high-lights the importance of using accurate extinction valueswhen estimating source distances. Given the typical un-certainties in the extinction towards black hole soft X-raytransients (Jonker & Nelemans 2004), we conclude thatthis issue is likely to affect the majority of such systems,rendering their distance estimates also uncertain.

3.2. Source luminosity

Our measured distance is significantly closer thanthe 3.5–4.0kpc commonly assumed for the source(e.g., Jonker & Nelemans 2004; Gallo et al. 2005;Bradley et al. 2007; Corbel et al. 2008), reducing itsluminosity by a factor of 2.2–2.8. The quiescent 0.3–10 keV unabsorbed flux of 1.08 × 10−12 erg cm−2 s−1,derived using a power-law model for the X-ray spectrum(Bradley et al. 2007), implies a quiescent luminosityof ∼ 7 × 1032 erg s−1. Makino (1989) measured amaximum flux of 17Crab with Ginga during the 1989outburst of V404 Cyg, implying a 1–70keV luminosityof 7.0 × 1038 erg s−1, which is of order 0.5LEdd for theblack hole mass of 12 ± 2M⊙ derived for V404 Cyg(Shahbaz et al. 1994). Thus the system was not super-Eddington during the 1989 outburst. Furthermore, inreducing the source radio and X-ray luminosities, ournew distance will reduce the scatter in the radio/X-ray

correlation (Gallo et al. 2003), bringing the pointsmeasured for V404 Cyg into better alignment with thoseof GX 339-4.

3.3. Peculiar velocity

With a more accurate distance and proper motion forthe source, and the updated Galactic rotation parameters(R⊙ = 8.4 kpc, Θ⊙ = 254km s−1) given by Reid et al.(2009) (but note McMillan & Binney 2009), we can re-visit the analysis of Miller-Jones et al. (2009) to derive amore accurate peculiar velocity of 39.9±5.5km s−1. Thiserror bar now includes the uncertainties in both the dis-tance and in all three space velocity components. Thisnew value is significantly lower than the best-fitting pe-culiar velocity of 64 km s−1 derived for a 4 kpc distance(Miller-Jones et al. 2009), and can easily be achieved viaa Blaauw kick, such that no asymmetric supernova kickis required. The component of the peculiar velocity inthe Galactic Plane is 39.6 km s−1, which, compared tothe expected velocity dispersion in the Galactic Planeof 18.9 km s−1(Mignard 2000) for the likely F0-F5 pro-genitor of the donor star, is a 2.1σ result, implying aprobability of only 0.038 that the peculiar velocity is aresult of Galactic velocity dispersion (for more details,see Miller-Jones et al. 2009). We therefore find it mostprobable that the peculiar velocity arises from a natal su-pernova kick, the magnitude of which is consistent withrecoil due to mass loss (a Blaauw kick), with any addi-tional asymmetric kick being small.

3.4. Size constraints

Miller-Jones et al. (2009) placed an upper limit of1.3mas on the source size, corresponding to a physicalsize of < 3AU at our new distance. While we did not re-solve the source, our highest-resolution data, taken witha global VLBI array at 22GHz, constrain the source sizeto < 0.6mas, a physical size of < 1.4AU. We can there-fore place a lower limit of 106 K on the brightness tem-perature of the source. From the high brightness tem-perature, the observed flat radio spectrum (Gallo et al.2005) characteristic of a steady, partially self-absorbedconical outflow (Blandford & Konigl 1979), and the lo-cation of the source on the radio/X-ray correlation ofGallo et al. (2003), at the high-luminosity end of whichjets have been directly resolved (Stirling et al. 2001), weinfer that the observed radio emission is likely to arisefrom compact, steady synchrotron-emitting jets.

Our upper limit on the jet size can be compared tothat of the resolved jets in Cygnus X-1 (Stirling et al.2001), with an angular size of 15mas at 8.4GHz. Sincethe jet size scales inversely with observing frequency(Blandford & Konigl 1979), this implies 5mas at 22GHz,for a physical scale of 10(d/2kpc)AU. Since Heinz (2006)

found the jet length should scale as L8/17ν where Lν is the

jet luminosity, then the difference in radio luminositiesof the two sources suggests a size scale of ∼ 1.5AU forV404 Cyg at 22GHz. Thus we are beginning to probethe angular scales on which the jets may be resolved.

The scatter in the residuals after our parallax fit(0.10mas in R.A. and 0.14mas in Dec.) can also helpconstrain the size and stability of the jets, since anymotion of unresolved emitting knots along the jet willchange the brightness distribution and hence the mea-sured centroid position. In particular, a comparison of

Orbital Period

Parallax

9

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.001

0.010

0.100

1.000

NF

e (p

ho

ton

s cm

-2 s

-1)

obs1obs2obs1obs2

(a)

1036 1037 1038

L(2-10 keV) (ergs s-1)

0.01

0.10

1.00

10.00

EW

(k

eV)

obs1obs2obs1obs2

(b)

1036 1037 1038

L(2-10 keV) (ergs s-1)

-2

-1

0

1

2obs1obs2obs1obs2

(c)

Fig. 3.— Panel (a) shows the line intensity of the narrow Fe K↵ line to the total flux measured between 2–10 keV. A tentative positivetrend is noted between these quantities. Panel (b) shows a negative correlation between the equivalent width of the narrow Fe K↵ and thetotal flux. The extremely high value of 1 keV in the lower flux states suggest that we are not observing the direct continuum, but the outerdisk may be obscuring our view. The resulting continuum is a combination of scattered and reflected light from the outer and inner disk,respectively. Panel (c) shows the phenomenological power-law spectral index versus the total flux. � < 1.4 can not be produced by a simpleComptonization model, and therefore absorption, scattering and reflection from the disk are likely altering our view of the continuum.

(a) (b)

Fig. 4.— Panel (a) corresponds to the 8th epoch in the second observation, and shows the ratio of the data to a phenomological power-lawcomponent. Clear absorption features are observed in all but the Fe K↵ and Fe K� lines. The absorption line width increases from MgXII to Fe XXVI. For contrast, panel (b) shows the same ions from the 6th epoch of the second observation. Only emission is detected inthis epoch.

Narrow Fe K alpha Line

Thermally driven wind

Higginbottom & Proga

He-like Ion Structure1s2s 1S 1s2p 1P 1s2s 3S 1s2p 3P

r f i 1s2s 3S is metastable

f/I density or radiation flux

r/(f+i) recombination vs collisional excitation

P-Cygni Profiles from Disk Winds

• HL Cma

• UV profiles of C IV

• Dwarf Nova in Outburst

• Lines vary with inclination, M-dot and velocity structure

• Mauche & Raymond