Radiative winds, accretion disks and massive stars physics using

Transcription

Stellar Winds & interferometry
Hot (massive) Stars
Philippe Stee
(help from D. Bonneau, A. Meilland, P.
Crowther, G. Weigelt & colleagues)
UNSA - CNRS
Outline
• Physics of massive hot stars
• Application to active hot stars (NLTE
physics)
• The SIMECA code
• Interferometry
• Results from Be stars, B[e] stars, HAe/Be
• Interacting binaries, Ups SgR, Eta Car
• Conclusions
ESF Exploratory Workshop: Extreme Laboratory Astrophysics
2
Diagnostics of mass-loss
• Hot star winds can be observed
through a variety of observations
– UV resonance lines
– Optical or IR emission lines
– Continuum excess in IR or radio
• A theory of radiatively driven has been
developed for hot stars which is
broadly in agreement with
observations.
3
Spectral lines from winds
• Spectral lines from winds can easily be
distinguished from photospheric lines because
of their large width or wavelength shift due to the
outflowing motion of the gas in the wind.
• Wind lines can appear in absorption, emission or
a combination of the two (P Cygni profile). For
hot stars, two line formation processes tend to
dominate.
4
Formation of Spectral lines
• Line scattering – If a photon emitted by the photosphere
is absorbed by an atom, it causes an electron of the
atom to be excited. Very quickly the photon is re-emitted
by spontaneous emission. So it appears the photon was
only scattered in another direction. If the line transition is
from the ground state, the line is a resonance line and
the scattering is called resonance scattering. Most P
Cygni profiles are formed from resonance scattering.
• Line emission by recombination – If an ion in a stellar
wind collides with an electron it can recombine. The
most likely recombination is directly to the ground state,
however it may recombine to an excited state. The
resulting excited ion may then cascade downwards by
photo-deexcitation. Each de-excitation results in the
emision of a line photon. This process is responsible for
Hα emission in hot stars.
5
P Cygni profiles
The most sensitive indicators of mass loss from hot stars
are resonance lines from abundant ions, such as CIV
1550A in O stars, which generally have large oscillator
strengths.
If the column density of the absorbing ions in the wind is
relatively small, they can produce an absorption line that
shows the Doppler shift (blue shifted since material is
moving towards the observer)
If the column density of the absorbing ion is large, the
blueshifted absorption is combined with emission that is
symmetric about the line centre (from a halo of gas
surrounding the stellar disk), producing a blue-shifted
absorption component plus a red-shifted emission
component .
6
Formation of P Cygni profile
7
Wind velocities
Wind velocities of early-type
stars are directly measured
from so-called `black’ troughs
of blue-shifted saturated P
Cygni profiles – observed to
range from several hundred to
several thousand km/s
8
Far-UV spectroscopy
Far-UV FUSE atlas of OB stars in Magellanic Clouds
9
Emission lines
Optical lines appear in
emission for only very
extreme winds – generally
exclusively Hα at 6562A
may be used for OB
supergiants.
Strength provides information
on dM/dt since Hα is a
recombination line (since
proportional to ρ2)
Lines are narrower than UV
resonance lines since Hα
formed very close to star
where density is high.
Hα fit to LMC O4If star –
dM/dt=6, 8.5, 11x10-6 Mo/yr
10
IR and radio excess
Stars with an ionized stellar wind emit an excess of
continuum emission at long wavelengths, due to freefree `thermal’ emission from the wind with fν ∝λ-0.6 which
falls off much slower than the photosphere (fν ∝λ-2). The
f-f emission depends on the density and temperature
structure of the winds. If the wind velocity is known, the
radio flux provides information on the mass-loss rates
via
dM/dt ∝ v∞fν0.75 d1.5 λ-0.25
This technique is limited to relatively nearby (few kpc)
stars with strong winds due to poor radio sensitivity
since the free-free emission is very weak.
Mass loss rates of O supergiants are of order 10-6 Mo/yr
(compare with 10-14 Mo/yr for Solar Wind), rather lower
for O dwarfs.
11
Excess emission from wind
Here we show the
spectral energy
distribution of a hot
star with a strong
wind. The dashed
line is that expected
for a hydrostatic
photosphere without
a wind – the grey
region is the IR and
radio free-free
excess from the
wind. The excess is
much weaker for
most hot stars.
12
Radius (τ≈2/3) increases with λ
Since the free-free opacity
increases as λ2, the optical
depth along a line-of-sight
into the hot star wind also
increases with λ2 as does
the effective radius of the
star, so the `radio
photosphere’ corresponds
to ≈102 R*
13
What drives hot star winds?
• How are hot star winds driven? The dominant continuum
opacity source in O stars is electron scattering. Does this
drive the wind? If the force from electron scattering were
to exceed gravity (known as the Eddington limit) the
surface of a star could not remain bound and the star
would blow itself apart.
• Instead, stellar winds are rather stable, using electrons
bound in atoms to absorb the radiation: Radiation
pressure is transferred to the wind material via spectral
lines, which are plentiful in the UV. Bound e- provide
much more opacity than free e- (e.g. opacity from CIV
1550 exceeds e.s. by 106!)
• Hot stars, unlike the Solar Wind, have plenty of line
opacity in the ultraviolet where most of the photospheric
radiation is. This combination allows for efficient driving
of winds in hot stars by radiation pressure.
14
Doppler shift
• The large radiation force on ions due to their spectral lines
would not be efficient in driving a wind if it were not for the
Doppler effect.
• In a static atmosphere with strong line-absorption, the
radiation from the photosphere will be absorbed or scattered
in the lower layers of the atmosphere. The outer layers will
not receive direct radiation from the photosphere, so the
radiative acceleration in the outer layers is strongly
diminished.
• However, if the outer layers are moving outwards, there is a
velocity gradient, allowing the atoms in the atmosphere to
see the radiation from the photosphere as redshifted (it
appears to be receeding as viewed from gas in the
expanding atmosphere). The Doppler shift allows the atoms
to absorb undiminished continuum photons in their line
transitions.
15
Line-driving
• Radiative acceleration due to spectral lines in the
atmospheres of hot stars (main sequence OB stars,
OBA supergiants, central stars of Planetary
Nebulae, WDs) is very efficient for driving a stellar
wind.
• This radiative acceleration in hot star winds is
provided by the absorption and re-emission of UV
photons in ions of abundant elements (CNO, Fepeak) in the Lyman continuum (λ<912A).
16
Example
• The resonance line of NIV is at 765A, so an ion that absorbs
such a photon will increase its velocity by hν/mc=37 cm/s. To
accelerate a single N3+ ion to 2000km/s requires 5x106
absorptions. In fact, since the wind is a plasma, the
momentum gained by the N3+ ion is shared with all
constituents in the wind (via interactions with surrounding
protons, ions, e-, due to the electric charge of the ion).
• Ions which provide the radiative acceleration constitute about
only 10-5 of all ions by number (since H and He contribute
negligibly to the acceleration since fully ionized) so if a typical
ion increases its velocity by 20cm/s, the effective increase
per absorption is 2x10-3 cm/s. To accelerate the wind to
2000km/s requires 1011 absorptions!
• The terminal velocity is reached in a few stellar radii, so the
time to accelerate the gas is 3R*/v∞=104s if R*=10Rsun so ions
that provide the acceleration have to (and do!) absorb 107
ph/s (typical of strong lines with large oscillator strengths)
ESF Exploratory Workshop: Extreme Laboratory
Astrophysics
17
CAK theory
The theory of radiatively driven winds was developed by Castor
et al. (1975, CAK), with the radiative acceleration dependent on
the fraction of optically thick lines (parameter α), the number of
strong lines (parameter k) and ionization (δ). Using these
parameters, it was found that
⎛ α ⎞
dM
1 /(α −δ )
and
v
∝
⎟vescape
⎜
∞
∝k
⎝1−α ⎠
dt
i.e. the mass-loss rate is predicted to scale with the number of
strong lines, and the wind velocity is predicted to scale with
escape velocity,
vescape = 2 g eff R
where the effective gravity is the stellar gravity corrected for the
reducing effect of radiative pressure via Γ (related to M and R)
GM
g eff = 2 (1 − Γ)
R
18
Sum of line driving
The total radiative acceleration (and hence
α,k,δ) can be calculated by summing the
contributions of all possible lines for all
elements. This is rather difficult, but was first
carried out by Abbott (1982). There are
wavelength regions where few lines
contribute to the acceleration and others
which are crowded with lines (e.g.
300<λ<600A in O stars).
Typically k>0.1, α=0.65, δ=0.1 from summing
all ions of all elements in O stars, predicting
wind properties that are in reasonable
agreement with observed wind properties.
It has recently been found that CNO
elements dictate the line driving in the outer
wind (hence v∞) whilst Fe-peak elements
control the inner wind (hence dM/dt)
19
Metallicity influence of winds
For O and early B stars
¾ dM/dt ∝ (Z/Zo)0.85 predicted
from radiatively driven wind.
Observationally, stars in metal
poor SMC galaxy do possess
weaker winds than Galactic
counterparts.
¾ v∞ ∝ (Z/Z)0.1 predicted from
theory, whilst observations
show slower winds in SMC O
dwarfs (thick lines) than
Galactic stars (thin lines) from
CIV 1550A & NV 1240A
ultraviolet resonance lines.
20
Multiple scattering
The standard theory of line driving
assumes that photons can be
scattered only once in the wind
which is a reasonable assumption
for normal O stars.
Line driving in WR stars is still
controversial, since the strength of
their winds appears to exceed the
single scattering limit. The
absorption by photons in different
spectral lines is called multiple
scattering. The process of multiple
scattering is shown schematically
here.Each scattering occurs in a
different specral line, successive
scatterings occur at lower energy
(longer wavelength).
Astrophysics
21
Applications to active hot stars
Be stars and other friends
22
Open questions
• Origin of the Be phenomenon:
– Why some hot stars are forming disks and some
others not ?
– What is the effect of the rotation ?
– What is the effect of the magnetic field ?
– What is the influence of stellar winds ?
– What is the importance of these disks on the
stellar evolution ?
– What is the impact on the ISM ?
23
Active hot stars = NLTE physics
24
Non-LTE for hot stars
Radiation field is so
intense in hot stars
(O-type, OBA
supergiants, WDs)
that their popluations
are only weakly
dependent on local
(Te,Ne), consequently
LTE represents a poor
assumption.
Eddington limit: Radiation pressure equals gravity
25
Non-LTE in OB stars
• O and early B dwarfs possess intense radiation fields
in which LTE is invalid. Hydrostatic equilibrium is
invalid in OBA supergiants – their tenuous
atmospheres lead to a drop in the line source function
below Planckian value.
• In O stars, LTE profiles are much too small.
Departures from LTE make He I and He II lines much
stronger.
• For B stars, in the blue-violet spectra of B stars, some
He I lines are formed in LTE, however red and IR lines
are not collision dominated, instead photoionizationrecombination processes dominate, so non-LTE is
necessary.
• In A supergiants, reliable metal abundance
determinations require non-LTE treatment – lines
become stronger in non-LTE with corrections of up to
factor of 10 for strong lines.
26
Hydrogen lines in βOri (B8Ia)
LTE (dash) and non-LTE (solid)
lines agree for Balmer lines (left),
but LTE lines too weak for
Paschen lines (below) in β Ori
27
The SIMECA code
SIMulation Etoiles Chaudes
Actives
(NLTE code)
28
SIMECA : Basic Scheme
Envelope and Stellar Physical Parameters :
Temperature = f(θ)
Stellar Radius = f(θ)
Photospheric Density = f(θ)
Stellar rotational velocity
Inclination
Velocity at the photophere
Equatorial terminal velocity
Polar terminal velocity
Polar Mass flux
H/H+He
Input Parameters :
Free Parameters
Hydro Code (CAK) :
ρ, Vr, VФ, T
Statistical Equil.
n1,…,n7,ne (LTE)
m1: variation of the mass flux
m2: variation of the terminal velocity
C1 : equatorial/polar mass flux ratio
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
29
Based on hydrodynamical equations (CAK)
-Continuity Equation
-Mass flux conservation
-Perfect gas state equation
Input Parameters :
Hydro code (CAK) :
ρ, Vr, VФ, T
Statistical Equil.
n1,…,n7,ne (LTE)
Hypothesis:
-Axial Symmetry (can be overcome)
-Stationarity
Temperature independent of stellar latitude
-Velocities independent of stellar latitude
-Radiative Pressure due to line and continuum
We obtain in the envelope the following distributions:
-Density
-Velocity fields (expansion+rotation)
-Temperature (as a function of r)
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
30
We start with the LTE population levels (1 to 7 levels + continuum)
We solve the statistical population equations in the Sobolev
approximation :
∞
⎛ i−1
⎞
n i ⎜ ∑ Aik b ki + Bicric ⎟ = ∑ n k Akib ik + n 2eC i (Te )
⎝ k=1
⎠ k= i+1
Input Parameters :
Hydro code (CAK) :
ρ, Vr, VФ, T
Statistical Equil.
n1,…,n7,ne (LTE)
Aik, Bic et Ci : Absorption coefficients, spontaneous emission and
recombination
βik : Escape probability (function of the velocity gradient)
We calculate the level populations by iterations until we reach
convergence in a 410x90x72 box.
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
31
Transfert Equation :
dIυ
= −κ υ Iυ + ευ
dz
We calculate τ by integration: dτ =-κ . dz (along the line of sight)
Input Parameters :
In the continuum :
-Envelope Opacity : free-free emission and electronic diffusion
-Envelope Emission : free-free and free-bound emission
Hydro Code (CAK) :
ρ, Vr, VФ, T
In the lines :
-Calculate κ et ε for the given transition
- Sobolev Approximation
Statistical Equil.
n1,…,n7,ne (LTE)
Intensity as a function of spatial coordinates:
(in a plane perpendicular to the line of sight)
For a given transition (line) or as a function of the wavelength (continuum)
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
32
In a spectral line :
Input Parameters :
Compute the iso-velocity regions Ù Doppler Effect
- Integrate these regions
Line Profiles
- Integrate within a spectral bandwith
Intensity maps
In the continuum :
Hydro code (CAK) :
ρ, Vr, VФ, T
Statistical Equil.
n1,…,n7,ne (LTE)
-Spatial Integration
Spectral Energy Distribution (SED)
-Spectral Integration
Intensity maps in the continuum
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
33
Input Parameters :
Hydro code (CAK) :
ρ, Vr, VФ, T
Statistical Equil.
n1,…,n7,ne (LTE)
n1,..,n7,ne (NLTE)
Radiative Transfert
Continuum
Radiative Transfert
Lines
Radiative Transfert
Continuum
SED
Line Profiles
Intensity Maps
in Contiuum
Intensity maps
In Lines
34
Interferometry
35
Interferometry
Pb : How to obtain information on the
brightness spatial distribution of an object
?
=>Direct observation limited by the spatial
resolution of the telescope
•Spatial resolution of a monolithic telescope
:
Image of a point source=> Airy disk
(R=1.22λ\D)
Resolution limit:
Minimal angle to separate 2 points on the
source(≈1,6λ\D)
θ (mas)= 250 λ (micron) / D (m)
• In the visible:
1.5’’
=> D≈10cm
1.5 mas
=> D≈100m
(3 m on the Moon)
0.005 mas => D≈30000m
(1 cm on the Moon)
α=λ\D
Non résolu
α=1,6λ\D
limite de résolution
α=2.44λ\D
résolu
36
Interferometry
Interferometry with 2 telescopes
Two telescopes with a diameter D
separated by a baseline B
Similar to the Young’s fringes
objets
experiment:
Light interference from a single source
⇒Fringes in the Airy disk:
Diameter : d = 2.44 λ \ D
Interfringe : i = λ \ B
Images dans un télescope monolithique
Interferometer Ù Telescope with a
diameter B + «mask with 2 holes with a
diameter D»
⇒Spatial resolution equivalent to a
monolitique telescope of diameter B.
Images dans un interféromètre à 2 télescopes
Spatial resolution of an
37
interferometer: ESF Exploratory Workshop: Extreme Laboratory Astrophysics
θ(mas)= 250 λ (micron) \ B (m)
Visibility Function
Interferometry
Visibility (V) = Fringes contrast ( between 0 and 1)
• Point source: V = 1
Fully resolved source: V=0 (no more
finges !)
• Extended sources: Van-Cittert et Zernike theorem =>V equal to
the
modulus
of the
object
at (u,v)= B\λ
Exemple
: two
stars
withFTdifferent
diameters.
D1= 10 mas et D2 = 4 mas
Limb darkned disk
Visibility as a function of baseline
between 0 and 100 meters
1 m : V1² = 1 et V2² = 1 (object not
resolved
50 m : V1² = 0 et V2² = 0.6 ( object 1
fully resolved)
38
The Very Large Telescope Interferometer
2 instruments available
MIDI
8-13μm
2 telescopes
Modulus and
Differential phase
R=200
θmin=10mas
AMBER
1.2-2.3μm (JHK)
3 telescopes
Modulus,
Differential phase,
and phase closure
R=35,1500,10000
θmin=2mas
uv coverage after 8 hour
observation with all UTs
(object at -15ο)
Airy disk
of UT
Resulting PSF is the Fourier transform of
the visibilities at λ = 2.2μm (K-band)
40
First fringes with the UTs
(Oct 2001)
UT1
UT3
Fringes on Achernar
(100-m baseline)
41
42
Spectro-interferometry
(Doppler Effect)
In the whole line
Geometry
Spectro-interferometry
(Doppler Effect)
In the whole line
Narrow spectral bandwith within the line
Variation of the visibility modulus and
phase as a function of wavelength
Geometry + Kinematics
Geometry
Expansion/rotation, rotational law,
inhomogeneities…
Be Stars:
α Arae
45
α Arae (VLTI/AMBER)
Intensity map computed with
SIMECA (continuum @ 2 μm)
46
What is the rotation law within
the circumstellar disk ?
Line profile and modulus
of the visibility
Differential phases
47
Keplerian ! ( Q. depuis 1866…)
Vrot
Vθ ≈ β
r
Keplerian β = 0.5
Angular Cons. Mt β =1
“First direct detection of a Keplerian rotating disk around the Be star α Arae using the
VLTI/AMBER instrument” Meilland, A., Stee, Ph. et al. 2006, A&A, astro-ph/0606404
48
And for a B[e] star ?
49
VLTI-AMBER : B[e] CPD 57-2874
Differential phase
UT2-UT3
λ (microns)
50
B[e] star: CPD 57-2874: clearly
an outflow !
51
κ CMa
An asymmetry detected
in the disk of KCMa
“An asymmetry detected in the disk of KCMa
with the VLTI/AMBER” Meilland, A., Millour, F..,
Stee, Ph. et al. 2006, A&A, accepté
52
Stellar wind and Be phenomenon
• Achernar
“The spinning Top Be star Achernar from
VLTI/VINCI”
Domiciano de souza et al. 2003, A&A, 407, L47
“The polar wind of the fast rotating Be star Achernar
VINCI/VLTI interferometric observations of an
elongated polar envelope”, Kervella, P. & Domiciano,
A. 2006, A&A, 453, 1059
¾Wind flux ≅ 5% of the stellar flux
¾Elongation > 10 stellar radii
53
Is it possible to interpret Domiciano et al. data
with another scenario?
≈
Less flattened star
Flattened star
+
Polar Jets
+
Small equatorial disk
+
Polar Jets
Disk and wind evolution of
Achernar: the breaking of the
fellowship
Kanaan, S., Meilland, A., Stee Ph. et al. 2008,
A&A, 486, 785
55
κ CMa
Kinematics
Meilland et al. 2006 A&A
No expansion (VΦ >> Vr )
Achernar
Kervella et al. 2005 A&A
Vinicius et al. 2005 A&A
non-Keplerian rotation (β =0.3±0.1)
Critical rotation
Rotation sub-critical Vrot~0.52Vc
No disk during VINCI
observations in 2002
Inhomogeneity ?
Disk between 1991 and 1998?
Astrophysics
56
Disk Formation and Dissipation
Be Stars : One Ring to rule them all?
Meilland et al. 2006, A&A, 455, 953
Ring
vs
Mass-Flux variation
Study the variation of observables during the
Disk dissipation
Visibilities
Line profiles
SED
Amplitude and position
Of the visibility second lobe
Double-pics separation
Time of the dissipation
Astrophysics
57
disk formation and Dissipation
δ Sco
Miroshnichenko et al. 2003 A&A 408,305
Growing disk till 2000 (Periastron)
Rdisk(2003)~10R*
Vr~0.4kms-1
Keplerian ?
Multiple outbursts?
ESF Exploratory Workshop: Extreme
Laboratory Astrophysics
58
HAe/Be star: MWC 297
“Disk and wind interactions in the young stellar object MWC 297 spatially resolved with
AMBER/VLTI” Malbet F., Benisty, M. et al. 2006, A&A, astro-ph/0510350
59
Interacting binaries:
β Lyrae & Ups Sgr
ESF Exploratory Workshop: Extreme
Laboratory Astrophysics
60
Massive stars in interacting binaries
υ Sgr
β Lyrae
Massive interactive binary systems
¾ mass transfer and mass loss
¾ Complex circumstellar environment, rich in hot gas and dust
system with exchange of mass (M ~ 10-20 M)
a donor star losing mass towards a star hidden in an accretion disc or a circumbinary
structure (β Lyrae et υ Sgr )
collaboration with the czech group of the Ondrejov observatory
61
β Lyrae (HD 174638, δ = +33°, V = 3.4, d ≈ 270 pc)
¾ system in 1er phase of mass transfer
variable composite spectrum
moderate IR excess ⇔ no dusty environment
¾ photometric and spectroscopic binary: P = 12.9 d, dP/dt = -19 s/y i ≈ 86°
¾ 1994 international campaign
Spectroscopy + photometry
Spectro-Interferometry with GI2T
Ground baseN-S 51 m
λ = 654 – 677 nm R = 5000
Hα
HeI 6678
star 1 (gainer)
star 2 (donor) accretion disk B2 V ≈ 13 M
B6-8 II ≈ 3 M
gas stream
localisation de l’absorption dans Hα
calibrated visibilities
whatever the orbital phase
¾ ~ E-W unresolved source in the continuum
¾ ~ N-S resolved source in the Hα emission
¾ Jet Like structure in β Lyrae ? (Harmanec et al., 1996)
62
A new vision of β Lyrae
¾ Polarimetric observations:
orbital axis θ ≈ 160° (Rudy, 1979)
UV line polarization θ ≈ 162° (Nordsieck et al., 1995)
¾ radio observations:
resolved source with MERLIN array at ν = 5 Ghz (λ = 6 cm)
size ~ 60 x 47 mas à θ ≈ 157° (Umana et al., 2000)
A = 58.5 R
¾ Light curves and UV spectrum modeling.
(Linnel et al., 1998)
¾ 3-D gas dynamical simulations of mass transfer.
⇔ a ≈ 1 mas
jet structure ?
Star2 (donor)
B 6-8 II ∼ 3 M
accretion disc
(Bisikalo et al., 2000)
¾ Disentangling of donor and accretion disc spectra.
No strong dependence of Hα emission during the
orbital cycle.
(Ak et al., 2007)
Star1 (gainer)
B 2 V? ∼ 13 M
gas stream
scattering envelope ?
Localization of the Hα and He I emissions
Mass transfer and mass loss study with CHARA
¾ association of observations with high angular and spectroscopic resolution
¾ to solve the ambiguities of the interpretation of the spectro-photometric data
Visible (VEGA-CHARA) ⇒ origin of the Hα emission, resolution of the binary
Near IR (CHARA, AMBER-VLTI) ⇒ free-free emission and Brγ emission
63
β Lyrae observed with VEGA-CHARA ?
200
CHARA baselines
E1
North (m)
100
N
W1
0
E2
W2
E
Base
B (m)
Az(°)
S1-S2
34,0
170.3
E1-E2
65,9
56.6
W1-W2
107,9
99,1
-100
S2
S1
-200
-200
-100
0
100
200
East (m)
W1-W2 baseline
continuum
Hα emission
+
VEGA
λ = [around Hα]
low resolution
(λ / Δλ = 1500)
¾ a toy model for β Lyrae
0.4
0.30
At maximum elongation
Continuum
74°
Hα emission N 0.85 mas
N
1 mas
74°
E
E
0.5 mas
0.85 mas
V2
V2
~ 0.0
0.2 mas 1 mas
2.0 mas
time
~ 0.0
time
0.2 mas
Fdonor = 0.62, Fdisc = 0.38
Fbin = 0.18, Fjet = 0.82
at ϕorb = 0.25 and 0.75, using aperture synthesis effect:
¾ the "donor-accretion" binary must be resolved with E1-E2 and W1-W2
¾ the Hα emission source must be resolved
64
β Lyrae observed with VEGA-CHARA ?
N
CHARA WI-W2 baseline
+
VEGA
λ = [around Hα]
medium resolution (λ / Δλ = 5000)
Hα photocenter
Continuum photocenter
V peak
74°
HA = 0
≈25°
Hα
W1-W2 baseline
≈ 0,38 mas
≈ 0,85 mas
≈ 1 mas
~0,4 mas
R peak
continuum
Photocenter location depends of the central wavelength and of the flux ratio of the different emitting regions.
In Hα line, observed light come from the bulk of the emission in addition to subjacent continuum.
The Interferometric Differential Imaging technique (Vakili et al., 1997) allows to measure the relative phase of
the fringe visibility and to determine the relative position of the emitting regions.
¾ At λ ≈ 656 nm, for the 107 m baseline, the fringe spacing is i ≈ 1.26 mas.
¾ photocenter separation ~ 0.4 mas ⇔ ~ 110° bump in the curve of the visibility phase across
the spectral line.
¾ refine the location and extension of the Hα emission?
65
Massive stars in interacting binaries
Ups Sgr ( HD 181615, δ = - 16°, d ≈ 500 pc)
¾ brighter member of the type of extremely hydrogen-deficient binary stars (HdB stars)
¾ HdB are evolved binary systems in a second phase of mass transfer
SB2, P ≈ 137.9 d dP/dt = -24 s/y
IR excess ?
intense and variable Hα emission
strong IR excess ⇔ very dusty circumbinary environment
Hα
∼B2-B5 V pe
Teff ∼ 16000 K
M ∼ 4 M
∼A2 Ia shell
Teff ∼ 10500 K
M ∼ 2.5 M
Brγ, Hα emission ?
L1
Hα absorption
¾ possible accretion disc and jet-like structure ?
¾ Spiral nebulae ?
(Koubsky et al., 2006)
(Narai, 1967)
Mass transfer and mass loss study
¾ association of observations with high angular and spectroscopic resolution
¾ to raise the ambiguities of the interpretation of the spectro-photometric data
extension of the circumbinary envelope (VLTI- MIDI & AMBER)
origin of the Hα emission (VEGA on CHARA)
66
Eta Car (LBV)
67
The Luminous Blue Variable η Carinae and its Homunculus nebula: outburst in
1843 caused the Homunculus nebula
η Carinae
HST image of the Homunculus nebula, Gull et al. 2006
68
• Extreme Luminous Blue Variable with spectacular outbursts
• Eta Carinae‘s mass: 70 to 100 solar masses
• Dense aspherical stellar wind: diameter ~4 mas ~ 9 AU diameter
• High density of the stellar wind Æ star is not visible
• WR binary companion (P~5.5 yr), spectroscopic events …
• Observations: visibilities, differential & closure phases
• Resolution of η Car's aspheric wind region: continuum, Br γ & He I; interpretation
• see Weigelt et al., A&A 464, 87 (2007)
69
VLTI-AMBER spectro-interferometry
of η Car’s stellar wind
Artist‘s view
HST image of the Homunculus nebula
70
The inner 1 arcsec: speckle
objects, primary wind, binary,
& wind-wind interaction zone
Hot
companion
(P~5.5 yr)
Primary
Speckle objects: ejecta excited by
Eta Car
(bispectrum speckle
interferometry, ESO 3.6 m, 2008)
η Car AMBER
interferograms recorded
with spectral resolution
12000 and the
fringe tracker FINITO:
- 3 ATs, DIT 2 s, 2008
- Broad line: wind (~400 km/s)
- Narrow l.: speckle ejecta (50)
- Telluric lines
- Br γ 2.17: primary star wind
- He I: wind-wind interaction?
η Car interferograms
recorded with
AMBER and the
fringe tracker FINITO:
3 ATs
Discussion of the broad
& narrow lines
- 3 ATs, DIT 2 s
- Broad line: wind (~400 km/s)
- Narrow l.: speckle ejecta (50)
- FOV of AMBER observations:
ATs: 250 mas Æ narrow lines
caused by the speckle objects
UTs: 80 mas Æ no narrow lines
Speckle objects:
Bispectrum speckle
interferometry
Results: Elongated K-band shape:
Position angle of 120° & projected axis ratio of 1.2
In the continuum around the Br Gamma line,
we found an asymmetry towards PA ~120°
with a projected axis ratio of ~1.2.
This result confirms the earlier finding of van
Boekel et al. 2003 using VLTI/VINCI (K-band
continuum).
These observations support theoretical
studies which predict an enhanced mass loss
in polar direction for massive stars rotating
close to their critical rotation rate (Owocki et
al. 1996, von Zeipel 1924).
These models predict a higher wind speed &
density along the polar axis than in the
equatorial plane.
74
Summary:
(1)Resolution of η Car's optically thick, aspheric wind
region in the continuum & within Br γ & He I:
Spectral resolution 1500 and 12000; fringe tracker
observations with high SNR.
(2) 50% encircled-energy diameters (fit of Hillier et al.
model CLV shapes):
K cont.:
Br Gamma:
He I 2.06:
4.3 mas
9.6 mas
6.5 mas (5 mas = 11 AU)
(3) K-band elongation: PA~120° & projected axis ratio
of 1.2. This aspherical wind can be explained by
models for winds from luminous hot stars rotating
near their critical speed (e.g., Owocki et al. 1996).
The models predict a higher wind speed and density
along the polar axis than in the equatorial plane.
(4) We developed an aspherical stellar wind model
which can explain the spectra, visibilities, differential
& closure phases. Binary studies. Spectroscopic 75
event.
Conclusions:
9 Active hot stars: very nice laboratory to test NLTE physics
9 Interferometry: powerful tool to study this physics with
sub-mas spatial resolution.
9 We need NLTE & 3D models to constrain the data
9 Spectrally-resolved interferometry, i.e. with both
spatial & spectral resolution is a key to constrain these models
9 Multi-wavelengths studies are mendatory (physics = f(λ))
9 Laboratory astrophysics can strongly help us since same
processes can be studied in a « box », i.e. NLTE physics,
radiative transfer, 3D hydro, coupling hydro-rad…and developm
of codes also useful for the astrophysical community…
76

Radiative winds, accretion disks and massive stars physics using

Transcription

Similar documents

WT2000 - AEC Online

WT2000 - Alfa

Chris Geary - Harness Racing NSW

January 11 - Black Rider Press

The ESF in North Rhine-Westphalia

Sperreng Middle School

Registration form copy

Narrated by Whoopi Goldberg - American Museum of Natural History

“Baby Stars in the Universe”