Transcript The Peculiar Physics of Line-Driving

The Peculiar Physics of
Line-Driving
Stan Owocki
Bartol Research Institute
University of Delaware
Colloborators: Ken Gayley, U. Iowa
Joachim Puls, U. Munich
Steve Cranmer, CfA
Outline:
• Radiative force from free electron scattering
• Resonant amplification of line-scattering
• Doppler sweeping of thick lines
• CAK theory for steady, spherical wind
• Line-driven instability
• Multi-D winds with vector line-force:
• Winds from rotating stars
•Wind Compressed Disks (WCDs)
• WCD inhibition by nonradial line-force
• Spindown of wind rotation
• Colliding wind binaries
• Radiative braking
• Line-driven ablation
• Summary
Winds that Sail on Starlight
Stan Owocki
Bartol Research Institute
University of Delaware
Collaborators
K. Gayley, U. Iowa
J. Puls, U. Munich
D. Cohen, Bartol/UDel.
Outline







What is a Stellar Wind?
Intercepting Light's Momentum
Doppler Sweeping by Spectral Lines
CAK Model for Steady, Line-Driven Wind
Instability of Line-Driving
Simulating Wind Structure
Summary
What are stellar winds?
"A continuous outflow of mass from a star"
Solar Wind
*
*
*
*
*
Sun has a very hot (10^6 K) corona
High Pressure => expansion
Supersonic, v ~400-700 km/s ~ v_esc >> v_sound
But mass loss rate is tiny, 10^-14 Msun/yr
Implies sun will lose only 0.01% of mass in whole 10^10
yr life
Hot-Star Winds
*
Massive stars (M~10-50 Msun) are hot (T~few 10^4 K) and
luminous (L~10^5-10^6 Lsun)
*
Wind outflow diagnosed by asymmetric "P-Cygni" lines in UV
*
Show v ~ 1000-3000 km/s!
*
Much higher mass loss rates, up to 10^-4 Msun/yr
*
Affects:
Stellar Evolution
ISM energy and mass balance
Bubbles may even trigger star formation – "starbursts"
* Driven by radiation pressure, scattered in spectral lines
Castor, Abbott, Klein (1975; "CAK") developed basic
formalism
Formation of P-Cygni Line Profile
Intercepting Light's Momentum
* Light transports energy (& information)
* But it also has momentum , p=E/c
* Usually negligible, because speed c is so high.
* But becomes significant for very bright objects,
e.g. Lasers, Luminous stars, Quasars/AGNs
* Key question: how big is force vs., e.g. gravity?
* Expressed through electron scattering Eddington factor
L Th
gel 4r2c e
e L



GM
ggrav
4GMc
r2
* For sun,
5
O 2.710
* But for hot stars with
6
L105 10
LO M1050M
O

 ~ 1
Free Electron Scattering
Thompson Cross Section
Th
=
2/3 barn
Th
= 0.66e-24 cm2
Line Scattering
For High Quality Line Resonance:
Cross Section >> Electron Scattering
8
Q~t~10Hz
10s~10
4 7 3
Q~ZQ~1010~10
lines~QTh
15
glines~10 gel
7
3
lines~10el
1
3
}
iff
F Fthin
Doppler Shifting of Line-Absorption
in an Accelerating Stellar Wind
Line Scattering in an Expanding Wind
Optically thick line-force
Sobolev
approximation
independent of 
like inertia!
CAK model of steady-state wind
=>
+
inertia
gravity
line-force
Line-Driven Instability from
Perturbed Profile Doppler Shift
Time snapshot of a wind instability simulation
Velocity & Density vs. Height
CAK
Steady-State
1500
-10
-11
1000
-12
-13
500
-14
0
-15
0.0
0.5
Height (R * )
1.0
Snapshot of Velocity and Density
Plotted vs. Height and Mass
in a Time-Dependent Wind Simulation
Flow "Structure" on the Autobahn
Back Scattering from Multiple Line Resonances
in a Non-montonic Velocity Field
Formation of Black Troughs
in Saturated P-Cygni Line-Profiles
from Structured Stellar Winds
P-Cygni Profile Synthesized for a
Smooth (---) and Structured ( ___ )
Stellar Wind Models
Profile for smooth,
CAK Wind
Black Trough
from Structured Wind
Instability Models with Energy Equation
Feldmeier 1995
Computational Requirements for
Stellar Wind Simulations
CAK/Sobolev Models
* line-force computed from local density and velocity
gradient
* modest timing requirements, comparable to standard hydro
* allows for 2D (in principle even 3D) models, e.g. with
rotation, disks, even B-field
Instability Simulations
* Line-force requires nonlocal solution of radiation transfer,
in principle in hundreds of spectral lines of varying strength
* Current approximations use integral escape probabilities
* Requires computation of line optical depth
* Analytically averaged over power-law line ensemble
* But still requires nested integrations over
- angle (or ray)
- frequency
- depth
* Thus far most models artificially restricted to 1D
* Efforts toward 2D instability simulations
- 3-ray aligned grid
- Short characteristics
- 2nd order Sobolev (A. Feldmeier)
Ray Integration Grids for
2D Radiation Hydro Models
Co-Rotating Interaction Region Models
log(Density)
a.
b.
local
CAK
model
nonlocal
smooth
model
c.
nonlocal
structured
model
Ongoing Projects in Stellar Winds
Wind Compressed Disks
and a Wind Binaries
Radiative BrakingQuickTime™
in Colliding
GIF decompressor
are needed to see this picture.
O star
WR
* star
O star
*WR
star
Wind Rotation Spindown from
Azimuthal Line-Torque
a.
b.
-0.9
-90
-0.7
-70
-0.5
-50
-0.3
-30
-0.1
-10
g

(10 3 cm/s 2 )
[V (nrf)-V (wcd)]


*sin( )*r/R
eq
(km/s)
Summary
* Massive, hot, luminous stars have strong stellar winds
* Driven by line-scattering of stellar radiation
* Highly unstable, leading to:
- high speed rarefactions
- slower dense clumps
- separated by Reverse Shocks
* Non-monotonic velocity evident in UV line Black Troughs
* But reverse shocks produce few X-rays
* Ongoing problems
- 2D (& 3D) models of compressible turbulence
- explain X-ray scaling laws
- how small-scale instability affects global wind
structure, e.g. wind collisions, disks, etc.
- Role of line-driving in other luminous systems, e.g.
CV disks; AGNs/QSOs