Polarisation of WolfRayet and Other Hot, Massive Stars

1
Polarisation of Wolf-Rayet and Other Hot, Massive
Stars

• Nicole St-Louis
• Université de Montréal
• Centre de Recherche en Astrophysique du Québec

Astronomical Polarimetry 2008, La Malbaie, Québec
2008
2
Classical Wolf-Rayet Stars

• Wolf-Rayet stars are the descendants of the most
  massive O stars (gt25 M?)
• They are characterised by their strong winds
• (? 10-5 M?/yr, v8 ? 2000 kms-1)
• They represent a key evolutionary phase between
  the main sequence and the SN explosion
• They impart to the ISM a huge amount of energy
  and chemically enrich it with heavy elements
• They might even be the precursors to
  long-timescale GRBs

3
Continuum linear polarisation curvesof
Wolf-Rayet Stars
WR42
WR103
WR90
WR40
St-Louis, Moffat, Drissen, Robert, Bastien et al.
(gt1987)
4
Continuum linear polarisation curvesof
Wolf-Rayet Stars

• Massive-star winds, and in particular WR star
  winds, have ample free electrons to generate
  wavelength-independent linear polarisation.
• But

5
Spherically symetric wind -- single star
E
E
E
E
E
E
E
Net polarisation is zero
E
6
Continuum linear polarisation curvesof
Wolf-Rayet Stars

• Massive-star winds, and in particular WR star
  winds, have ample free electrons to generate
  wavelength-independent linear polarisation.
• But in spherically-symmetric winds, there should
  be no net polarisation.

• But some massive stars are binaries.

WR90
7
Spherically symetric wind --Binary
Companion
This angle is very small
The net polarization is
8
Continuum linear polarisation curvesof
Wolf-Rayet Stars
WC7O5-7
WR42

• The BME model works beautifully well and allows
  one to determine the orbital inclination which in
  turn can lead to the masses. Note of caution
  from Aspin, Simmon Brown (1981)

9
Mass-Loss Determination Methods
??
??2

• Resonnance line fits
• (Fullerton, Massa, Prinja 2006)
• Orbital period lenghtening due to a radially
  symetric mass-loss (Khaliullin 1974) (B)
• Photometry (atmospheric eclipses, e.g. Lamontagne
  et al. 1996)(B)
• Polarimetry (B)

• Recombination line fits (H?)
• Radio/IR free-free emission (Wright Barlow
  1975)

• H? fitting based on the technique developped by
  Puls et al. (1996)
• Improved by Markova et al. (2004) to include
  line blocking/blanketing

10
Polarimetry

• For WR stars, St-Louis et al. (1988)
• Ap is the semi-amplitude of the ellipse in the
  Q-U plane.
• a(R?) is the semi-major axis of the orbit
• v? is the terminal velocity of the WR wind
• i is the orbital inclination
• fc is the fraction of the total flux coming from
  the companion
• I is the numerical value of an intergral over
  angles. This depends on ? and on the starting
  radius

11
Polarimetry - WRO Systems
WR 42HD97152
St-Louis et al. (1987)
12
Comparison with other estimates
x IR (free-free) ? radio (free-free) Factor of
3 lower than other estimates in most cases
13
Polarimetry - the case of V444 Cygni
Robert et al. (1990)
14
Polarimetry - the case of V444 Cygni
From period change, Khaliullin et al. (1984)
Radio free-free emission estimates agree with
polarisation. IR less so
15
Polarimetry of O O systems
St-Louis Moffat (2008)

• A litterature search yielded 6 OO binaries with
  published polarimatric binary curves. Mostly III
  and I but two O4f systems as well.
• Here there are two winds The polarisation
  amplitude is the total value for both winds taken
  together if the winds are equal, it is doubled.
• Wind consists of H instead of He
• ?0.8
• In most systems, H? profiles have been found to
  be affected by flows related to colliding winds
  so the cant be used to estimate mass-loss rates

16
Results
i Luna (1988), ii Rudy Herman (1978), iii Lupie
Nordsieck (1987), iv Niemela et al. (1992), v
Morrell Niemela (1990)
17
The case of HD149404

• HD 14904 is the only star in our sample that we
  found to have other mass-loss rate
  determinations
• Polarimetry 3.1 x 10-6 M?yr-1
• Radio 6.2 x 10-6 M?yr-1
• PV resonnance lines 6.2 x 10-8 M?yr-1 (for
  qP41) (two stars?)
• Polarimetry seems to support a clumping factor of
  ? 3 but not 10-100!

18
Continuum linear polarisation curvesof
Wolf-Rayet Stars
WR40

• Single WR stars can still show linear
  polarimetric variability.
• The timescale is relatively short.
• We now know that winds in WR stars are clumped.
• Small subpeaks are superposed on top of strong
  emission lines and move from line center towards
  line edges on relatively short timescales
• This can easily produce continuum linear
  polarimetric variability as observed.

19
Continuum linear polarisation curvesof
Wolf-Rayet Stars

• One hint was found in Drissen et al. (1987). A
  trend for slower winds to show a smaller linear
  polarisation scatter.
• The interpretation was that perhaps blobs formed
  in fast winds have more difficulty to survive
  they act as an homogenezing agent.
• A model of polarisation from blob ejection was
  devised by Davies, Vink Oudmaijer (2008). (see
  also a previous model by Li et al. 2000)

20
Clumping-induced polarimetric variability from
winds (Davies et al. 08)

• They first calculated P for one clump. Mass and
  angular sizes of clumps are conserved.
• They used a ? law to describe the movement of the
  clump
• Then they produced many randomly ejected blobs
  and added their Q and Us.

• They examined the effect of varying many input
  parameters such as ejection rate, mass-loss rate,
  clump size, radius, etc

21
Clumping-induced polarimetric variability from
winds (Davies et al. 08)

• ltPgt and ?(P) increase with mass-loss rate
• The increasing ?(P) with decreasing v8 is
  explained by the fact that blobs spend more time
  in the wind

• The observed ltPgt and ?(P) has two possible
  explanations
• Because of the timescale of the variability,
  statistical deviations from spherical symmetry is
  prefered over a small number of blobs

22
Spectropolarimetry

• If there is a global asymmetry in the wind, then
  the so-called line effect is observed
• The global asymmetry causes the continuum
  polarisation from electron scattering to be
  polarised.
• If the lines are formed by recombination, then
  they should not be polarised.
• But the unpolarised line photons can subsequently
  be scattered and thus the lines can be polarised
• This can be (as has been) used to detect
  non-spherically symmetric winds.

23
McLean et al. (1979)
Spectral resolution 50 Å
24
HeII ??4686 must have a scattering component
which increases its polarisation
25
Schulte-Ladbeck, Nordsieck et al. (1991)
IS

• Wind is not spherically symmetric
• Cont pol. due to e scattering (inclined disk
  model) as continuum polarisation is grey
• Lines are polarised due to e-scattering
• Confirm ionisation stratification

Spectral resolution 34 Å
Schulte-Ladbeck et al. (1992) carried out a
similar study and obtained similar results for
WR134 except that continuum polarisation rises to
the UV
26
Harries, Hillier Howarth (1998)
134,40

• Spectropolarimetric survey of 16 northern WR
  stars.
• Data from the WHT. Resolution of ? 3 Å.
• Four stars were found to show line
  de-polarisation (WR 134, WR 137, WR 139, WR 141
  --binaries).
• A statistical analysis shows that this fraction
  is best reproduced if only 20 of stars have an
  intrinsic pol. gt 0.3 (equatorpole23)

136
6
137
16

• Combining their data with those of Schulte
  Ladbeck (1994), they found that stars with the
  highest mass-loss rates are those found to show
  the line effect, i.e. to have flattened winds.

27
Vink (2007)

• Carried out a Spectropolarimetric survey of 13 WR
  stars in the LMC to search for a different
  behaviour in a low-metalicity environment
  (possible GRB progenitors).
• Data from the VLT-FORS1. Resolution of ? 3 Å.
• Two stars were found to show line de-polarisation
  (BAT22 BAT33)
• The fraction is similar to that of galactic
  (?15).
• If metallicity is important in contributing to
  reduce the angular momentum loss from WR winds,
  the threshold is below that of the LMC (lt0.5 Z?)

28
Other stars

• 1) Harries, Howarth Evans (2002) -- 20 O
  Galactic supergiants
• 5 were found to show the line depolarisation
  effect (25)

• Davies, Oudmaijer Vink (2005) -- 14 Galactic
  and Magelanic Cloud LBV
• 50 (!!) are found to show the line effect (H?).
• 4 were observed during multi-epochs, 3 with
  random polarisation angles and 1 with a constant
  angle.
• They interpret this as due to the presence of
  strong clumps in the wind (consistent with their
  model)

29
ESPaDOnS data of WR6HD50896 De la Chevrotière,
St-Louis, Moffat (2008)
30
First detection of a magnetic field in a WR star
?100 Gauss To be modelled with Ignace Gayley
(2003) improvments to come
31
Summary
Polarisation

• Contributed to help us confirm that because winds
  are clumpy, mass-loss rates need to be revised
• Can allow us to determine the orbital inclination
  in a WRO binary system
• The short timescale of random polarisation
  variability tells us that the winds contain many
  many small blobs instead of a smooth component
  a few big blobs
• Tells us that only 15-20 of WR stars have
  flattened winds which is what we expect
  because they are not expected to have a fast
  rotation
• Will allow us to detect magnetic fields if there
  are any

32

• Merci !