Nonthermal emission from earlytype binaries Gregor Rauw Institut dAstrophysique et de Gophysique, Un

1
Non-thermal emission from early-type
binariesGregor RauwInstitut dAstrophysique
et de Géophysique, Université de LiègeAllée du 6
Août, Bât. B5c, B-4000 Liège (Sart-Tilman),
Belgium
2
Overview

• Early-type stars general properties
• Colliding winds in early-type binaries
• Synchrotron radio emission from early-type stars
• Are all n.-t. early-type stars binaries?
• High-energy non-thermal emission?
• The connection with unidentified ?-ray sources

3
Early-type stars general properties

• Spectral types O and Wolf-Rayet (WR)
• Massive objects
• 
  
• main sources of UV radiation in the ISM

4
Early-type stars general properties

• Stellar winds driven by radiation pressure
• huge mass-loss rates
• large velocities
• radiatively driven winds are unstable ? formation
  of shocks and structures (e.g. Dessart Owocki
  2003, AA 406, L1)

Intrinsic shocks can accelerate particles through
1st order Fermi mechanism (Chen White 1991, ApJ
366, 512)
5
Colliding winds in early-type binaries

• Winds of early-type stars in binary systems
  interact in strong hydrodynamical shocks
  post-shock gas heated to several million K.

Contact discontinuity cone with shape set
by wind momentum ratio

• Stevens et al. (1992, ApJ 386, 265)

6
Colliding winds in early-type binaries

• Observational signatures of wind interactions
  over a broad range of energies
• X-rays hot plasma in post-shock region produces
  an excess emission that displays phase-locked
  variations (e.g. Sana et al. 2004, MNRAS 350, 809)

7
Colliding winds in early-type binaries

• Observational signatures of wind interactions
  over a broad range of energies
• X-rays hot plasma in post-shock region produces
  an excess emission that displays phase-locked
  variations (e.g. Sana et al. 2004, MNRAS 350, 809)

8
Colliding winds in early-type binaries

• Observational signatures of wind interactions
  over a broad range of energies
• UV and optical when the post-shock gas cools
  efficiently, it can produce variable
  recombination emission lines (e.g. Gosset et al.
  2001, MNRAS 327, 435)

9
Colliding winds in early-type binaries

• Observational signatures of wind interactions
  over a broad range of energies
• UV and optical when the post-shock gas cools
  efficiently, it can produce variable
  recombination emission lines (e.g. Gosset et al.
  2001, MNRAS 327, 435)

10
Colliding winds in early-type binaries

• Observational signatures of wind interactions
  over a broad range of energies
• IR in high-density, efficiently cooling
  post-shock plasma of WC O systems, dust can
  form episodically (e.g. Williams 2002, ASP Conf.
  260, 311)

11
Colliding winds in early-type binaries

• Shock can be radiative (? 4) or adiabatic (? ?
  4) depending on the importance of radiative
  cooling

Wind-wind shocks can accelerate particles through
1st order Fermi mechanism (Eichler Usov 1993,
ApJ 402, 271)
12
Synchrotron radio emission from early-type stars

• The dense stellar winds of early-type stars
  produce an intense thermal (free-free) radio
  emission with a spectral index ? ? 0.6
• The winds are optically thick out to large radii
  at radio wavelengths
• For a typical O-star t 1 at about 50, 75 and
  175R _at_ 3.6, 6 and 20cm respectively

13
Synchrotron radio emission from early-type stars

• 25 40 of the massive stars within 2.2 kpc
  display a radio flux larger than expected from
  the free-free emission corresponding to their
  mass loss rates
• Their radio emission is often variable and has a
  negative spectral index a ? 0.0 (Bieging et al.
  1989, ApJ 340, 518)

14
Synchrotron radio emission from early-type stars

• 25 40 of the massive stars within 2.2 kpc
  display a radio flux larger than expected from
  the free-free emission corresponding to their
  mass loss rates
• Their radio emission is often variable and has a
  negative spectral index a ? 0.0 (Bieging et al.
  1989, ApJ 340, 518)

• Non-thermal (synchrotron) radio emission!
  

15
Synchrotron radio emission from early-type stars

• Synchrotron radio emission ? ? a magnetic field
  and a population of relativistic electrons in the
  winds of these objects
• Particle acceleration in stellar winds
• 1st order Fermi mechanism (Bell 1978, MNRAS 182,
  147 443) requires shocks OK, in single and
  binary stars
• ? relativistic electrons with power-law
  distribution of index n (?2)/(?-1)

16
Synchrotron radio emission from early-type stars

• Magnetic fields in early-type stars?
  So far, only two direct measurements
  360G (ß Cep, B1 IV) and 1100G (? Ori, O4-6V)
  (Donati et al. 2001, MNRAS 326, 1265 2002, MNRAS
  333, 55) In most
  cases upper limits of hundred Gauss Several
  mechanisms to produce a magnetic field have been
  proposed (see poster by De Becker et al.).
  

17
Synchrotron radio emission from early-type stars
18
Synchrotron radio emission from early-type stars
Variability expected in long-period, eccentric
binaries
19
Synchrotron radio emission from early-type stars

• WR140 (WC7 O5, period 7.9 yrs, e0.84)
  phase-locked radio variability emission
  increases and becomes n.-t. between f 0.55 and
  f 0.95

• due to variation of optical depth along the line
  of sight and eccentricity (White Becker 1995,
  ApJ 451, 352)

20
Synchrotron radio emission from early-type stars

• Cyg OB2 5 consisting of a Of Ofpe/WN9 binary
  (6.6 day period) early B star at 0.95
  elongated n.-t. radio emission resolved between
  the close binary and the third star (Contreras et
  al. 1997, ApJ 488, L153)

21
Synchrotron radio emission from early-type stars

• WR 147 (WN8 OB, visual binary) radio emission
  resolved by MERLIN into 2 components n.-t.
  emission between the 2 stars at a location
  consistent with colliding wind scenario (Williams
  et al. 1997, MNRAS 289, 10)

22
Are all n.-t. early-type stars binaries?

• Most of the n.-t. WR stars (7 out of 9 studied by
  Dougherty Williams 2000, MNRAS 319, 1005) and
  many of the OB stars are in fact either
  long-period spectroscopic, astrometric or visual
  colliding wind binaries
• ? could it be that non-thermal radio emission
  from early-type stars is only observable if it
  arises from a colliding wind zone well outside
  the huge radio photosphere?

23
Are all n.-t. early-type stars binaries?
WR-stars 16 confirmed n.-t. radio emitters
24
Are all n.-t. early-type stars binaries?
O-stars 11 confirmed n.-t. radio emitters
25
Are all n.-t. early-type stars binaries?

• More than 2/3 of the n.-t. WR and more than half
  of the n.-t. O stars are multiple objects.
• But does it mean that they are all binaries?
• binary fraction among 227 known Galactic WR
  stars 39 (van der Hucht 2001, New Ast. Rev. 45,
  135)
• binary fraction among Galactic O stars in open
  clusters 75 (Gies et al. 1998, ASP Conf. 131,
  382)
• current knowledge of multiplicity among
  early-type stars not sufficient, especially for
  moderately long period systems

26
Are all n.-t. early-type stars binaries?

• Situation less clear for O-stars than for
  WR-stars
• 9 Sgr (O4V) probable SB2 with a long period
  (more than 15 years?, Rauw et al. 2002, AA 394,
  993)
• HD93129A (O2If) 55 marcsec visual binary
  (Benaglia Koribalski 2004, AA 416, 171 Nelan
  et al. 2004, AJ, in press)
• Cyg OB2 8A (O6 O5.5) SB2 with 23.3-days
  period (De Becker et al. 2004, see poster
  outside)
• HD168112 (O5III) no RV variations, but variable
  X-ray emission (De Becker et al. 2004, AA, in
  press)

27
High-energy non-thermal emission

• Enormous flux of photospheric UV photons in the
  winds of early-type stars relativistic
  electrons ? inverse Compton scattering becomes a
  major energy loss mechanism and may produce a
  strong n.-t. X-ray and ?-ray emission power-law
  spectrum from keV to MeV energies (Pollock 1987,
  AA 171, 135 Chen White 1991, ApJ 366, 512
  ApJ 381, L63)
• ? typical Lorentz factors
  of 100 to 10 000 can produce IC X-ray and ?-ray
  emission

28
High-energy non-thermal emission

• Predicted IC luminosities 1033 1034 erg s-1
  (Chen White 1991, ApJ 381, L63 Benaglia et al.
  2001, AA 366, 605)
• BUT depends strongly on poorly constrained
  parameters magnetic field, distribution of
  relativistic electrons

29
High-energy non-thermal emission

• In single stars stellar winds have different
  optical depths in the radio and high-energy
  wavebands and the n.t. emissions arise from
  different locations in the wind and thus imply
  different populations of relativistic electrons

30
High-energy non-thermal emission

• In wide binary systems relativistic electrons
  accelerated in colliding wind zone ? n.t. radio
  and high-energy emissions arise from same
  population of relativistic electrons

31
Some predictions for INTEGRAL
32
High-energy non-thermal emission
Multiwavelength campaign to study n.-t.
early-type stars with XMM-Newton, INTEGRAL,
ground-based optical telescopes and the VLA
33
High-energy non-thermal emission

• 1st results of this campaign
• detection of hard X-ray tail in the XMM spectra
  of 9 Sgr (Rauw et al. 2002, AA 394, 993) and
  HD168112 (De Becker et al. 2004, AA in press)
  but most likely thermal (colliding winds?)
  emission

34
High-energy non-thermal emission

• 1st results of this campaign
• 9 Sgr probably SB2 binary with long period no RV
  variations seen for HD168112 Cyg OB2 8A SB2
  with period of 23.3 days (De Becker et al. 2004,
  see poster)
• Ongoing INTEGRAL observations of Cyg OB2 so far,
  80 ksec obtained, but no obvious detection yet!

35
The connection with unidentified ?-ray sources

• Unidentified EGRET sources correlated with OB
  associations and some extreme (n.-t.) O and WR
  stars (Romero et al. 1999, AA 348, 868)

• e.g. Cyg OB2 3 n.-t. radio
• O-stars (5, 8a and 9)
• inside the error box of
• 3EG J20334118 (Benaglia
• et al. 2001, AA 366, 605)

36
Other mechanisms contributing to ?-ray emission
from OB associations

• ?0 decay (EGRET energy range) due to
• relativistic protons injected by young stars,
  accelerated by SNRs and interacting with
  molecular clouds (SNOB scenario, Montmerle 1979,
  ApJ 231, 95) ? diffuse emission
• protons accelerated at the terminal shock (Cassé
  Paul 1980, ApJ 237, 236) or interface between
  winds of stars in open clusters (Manchanda et al.
  1996, AA 305, 457)

37
Other mechanisms contributing to ?-ray emission
from OB associations

• ?0 decay (EGRET energy range) due to
• protons accelerated by intrinsic instabilities in
  the winds of individual stars (White Chen 1992,
  ApJ 387, L81)
• bremsstrahlung in the dense winds of WR stars
  (Pollock 1987, AA 171, 135)

38
Summary and Conclusions

• The winds of some early-type stars harbour
  relativistic particles that may produce a
  detectable signature over a broad energy range
• Synchrotron radiation is observed in the radio
  emission of a number of WR and OB stars.
• Inverse Compton X-ray and ?-ray emission is
  expected to result from the interplay between the
  relativistic electrons and the strong stellar UV
  radiation field.

39
Summary and Conclusions

• There is evidence that most (if not all) of these
  stars are (colliding wind) binaries.
• Theoretical work is needed to better model the
  formation of non-thermal radiation under various
  circumstances and to constrain the unknown
  parameters that play a key role.
• This phenomenon might account for some of the yet
  unidentified EGRET sources correlated with OB
  associations.
• INTEGRAL and GLAST may help constraining the
  properties of these systems.