X-ray Emission from O Stars David Cohen Swarthmore College

Outline – focus on single O stars 1.Young OB stars produce strong hard X-rays in their magnetically channeled winds 2.After ~1 Myr X-ray emission is weaker and softer: embedded wind shocks in early O supergiants 3.X-ray line profiles provide evidence of low mass-loss rates 4.Wind-wind binaries will not be discussed

Young OB stars are very X-ray bright L x up to ~10 34 ergs s -1 X-ray temperatures: few up to 10+ keV (10s to 100+ million K) Orion; Chandra (Feigelson et al. 2002)

~7’ Orion Nebula Cluster: age ~ 1Myr; d ~ 450pc

Color coding of x-ray energy: 2.5keV Chandra ~10 6 seconds, sub-arcsec resolution

Brightest X-ray sources are OB stars  1 Ori C (O7 V)

Pup  1 Ori C Chandra grating spectra (R ~ > 300 km s -1 )  1 Ori C: hotter plasma, narrower emission lines  Pup (O4 If): cooler plasma, broad emission lines

 Pup  1 Ori C Si XIII Si XIV Mg XIMg XII H-like/He-like ratio is temperature sensitive

 Pup  1 Ori C Si XIII Si XIV The young O star –  1 Ori C – is hotter Mg XIMg XII

Differential emission measure (temperature distribution)  1 Ori C: peak near 30 million K evolved O stars, peak at a few million K

Dipole magnetic field (> 1 kG) measured on  1 Ori C Magnetic field obliquity,  ~ 45 o Wade et al. (2006)

temperatureemission measure MHD simulations of magnetically channeled wind Channeled collision is close to head-on - at km s -1 : T = K

Emission measure contour encloses T > 10 6 K

MHD simulations show multi-10 6 K plasma, moving slowly, ~1R * above photosphere contour encloses T > 10 6 K

Differential emission measure (temperature distribution) MHD simulation of  1 Ori C reproduces the observed differential emission measure

1000 km s -1 Young O stars have only modestly broad (few 100 km s -1 ) X-ray emission lines But mature O stars have broad (> 1000 km s -1 ), asymmetric emission lines A different mechanism is responsible  1 Ori C (O7 V)  Pup (O4 If)

1-D rad-hydro simulation of an O star wind Radiation line driving is inherently unstable: shock-heating and X-ray emission

continuum absorption in the bulk wind preferentially absorbs red shifted photons from the far side of the wind Contours of constant optical depth (observer is on the left) wavelength red blue brightness

The basic wind-profile model R o =1.5 R o =3 R o =10   =1,2,8 key parameters: R o &  * j ~  2 for r/R * > R o, = 0 otherwise

 * = 2.0 R o = 1.5  Pup: Fe XVII line at Å - Chandra

Onset of instability-induced shock structure: R o ~ R * = height of 0.5 R *

A factor of 4 reduction in mass-loss rate over the literature value of 6 X M sun /yr  ~ 70 cm 2 g 15 Å 1.5 X M sun /yr for  * = 2 Wind optical depth constrains mass-loss rate Waldron et al. (1998)

 Pup: Fe XVII line at Å - again best-fit model, with  * = 2, is preferred over the  * = 8 model with >99.999% confidence best-fit: low mass-loss rate literature mass-loss rate

The key parameter is the porosity length, h = (L 3 /ℓ 2 ) = ℓ/f The porosity associated with a distribution of optically thick clumps acts to reduce the effective opacity of the wind h=h’r/R * ℓ=0.1 Porosity reduces the effective wind optical depth once h becomes comparable to r/R *

h = (L 3 / ℓ 2 ) = ℓ /f

h’ = 0.5 h’ = 1.0 h’ = 2.0 h = h’r

h = (L 3 / ℓ 2 ) = ℓ /f f = 0.1 f = 0.2 f = 0.05

The optical depth integral is modified according to the clumping-induced effective opacity:

Porosity only affects line profiles if the porosity length exceeds the stellar radius

 * =8; h ∞ =3.3  * =2; h ∞ =0.0 The Chandra line profiles in  Pup can be fit by a porous model… …but huge porosity lengths are required and the fits tend to be worse

Two models from previous slide, but with perfect resolution  * =8; h ∞ =3.3  * =2; h ∞ =0.0

95% 68% Joint constraints on  * and h ∞ Even a model with h ∞ =1 only allows for a slightly larger  * and, hence, mass-loss rate h ∞ > 2.5 is required if you want to “rescue” the literature mass-loss rate

There is NO evidence from X-ray line profiles for significant wind porosity

This degree of porosity is not expected from the line-driven instability. The clumping in 2-D simulations (below) is on quite small scales. Dessart & Owocki 2003, A&A, 406, L1

The small-scale clumps affect density squared diagnostics: reduced mass-loss rates But they do not affect the X-ray profiles directly Dessart & Owocki 2003, A&A, 406, L1

Conclusions Many young (< 1Myr) O stars produce strong and hard X-ray emission in their magnetically channeled winds As they age, softer and weaker X-ray emission is produced via embedded wind shocks in a spherically expanding wind The X-rays themselves provide information about the wind conditions of O stars, including more evidence for reduced mass-loss rates

NGC 6611: ~5Myr Chandra images M17: ~0.5Myr soft medium hard

EXTRA SLIDES

OB stars are the brightest sources of stellar x-rays in young clusters - they irradiate their circumstellar environment, including nearby stars, disks, planetary atmospheres X-rays are highly penetrating, so they’re especially good for identifying and studying embedded sources Role of X-rays in OB Stars

Trace dynamic and energetic processes in outer atmospheres Trace evolution of magnetic fields Provide diagnostics of wind conditions (test theories, fundamental parameters) Influence wind ionization conditions Relation to diffuse x-rays from bubbles? Role of X-rays in OB Stars

2-D MHD simulation of  1 Ori C: density courtesy A. ud-Doula

2-D MHD simulation of  1 Ori C: temperature courtesy A. ud-Doula

2-D MHD simulation of  1 Ori C: speed courtesy A. ud-Doula

Overall X-ray flux synthesized from the same MHD simulation snapshot. The dip at oblique viewing angles is due to stellar occultation. Data from four different Chandra observations is superimposed. The amount of occultation seen at large viewing angles constrains the radii at which the x-ray emitting plasma exists

He-like f/i line ratios in O stars are diagnostics of source location f fi i

 Pup  1 Ori C (O7 V) Capella

But the emission lines are quite broad Ne X Ne IX Fe XVII  Pup (O4 I) 12 Å 15 Å Capella (G 5 III) - a coronal source of soft X-rays

Each individual line (here is Ne X Lyat Å) is significantly Doppler broadened and blue shifted  Pup (O4 I) Capella (G5 III) HWHM ~ 1000 km/s lab/rest wavelength unresolved at MEG resolution

The basic smooth wind model: for r>R o R o =1.5 R o =3 R o =10 =1,2,8=1,2,8 key parameters: R o &  * j   2 for r/R * > R o, = 0 otherwise

Highest S/N line in the  Pup Chandra spectrum o -v ∞ +v ∞ Fe Å  Fe +16 – neon-like; dominant stage of iron at T ~ 3 X 10 6 K in this coronal plasma 560 total counts note Poisson error bars

95% 90% 68% 1.5 <  * < 2.6 and 1.3 < R o < 1.7 1/R o

Marc’s images

M17: ~0.5Myr 4’ soft medium hard

Orion Nebula Cluster: ~1Myr 4’ soft medium hard Dashed arrows point to very early B stars.

Tr 14: ~ Myr 4’ soft medium hard

NGC 6611: ~5Myr 7’ soft medium hard

The first three clusters, the ~1Myr ones, have a mix of softer (and non-variable) and harder (variable) O star But the last one, the older, ~5Myr, NGC 6611 in M16, has only softer O stars The color coding is not identical in each frame; it’s adjusted to compensate for their different ISM column densities. But roughly, the RGB channels correspond to 2 keV. In many cases, B stars are as hard or harder (and brighter) than O stars. The very hardest stars are always flaring late-type T Tauri stars. What’s happened to the hard, variable O stars by 5 Myr?