X-ray emission from massive stars

home    research    teaching    personal    links    contacts
Stars across the HR diagram emit X-rays. The low- and solar-mass stars possess X-ray emitting coronae which are powered by an outer convection zone via magnetic fields. However, stars of spectral types earlier than A7 have no outer convection zone and normally no surface magnetic field. Therefore, the discovery of X-ray emission from OB-type stars was surprising (Seward et al. 1979), albeit Cassinelli & Olson (1979) had already attributed the ``superionization'' observed in the UV spectra of O-star winds to the presence of an internal  X-ray field. The ROSAT All Sky Survey provided firm evidence that X-rays are commonly emitted from massive-star winds. A rough correlation between X-ray and stellar bolometric luminosity has been empirically established (Berghoefer et al. 1996) .
X-rays can be detected from the regions on sky which cannot be observed in visual because of high interstellar absorption. X-ray emission is most importnant to understand stochastic non-statianary processes operating in stellar winds.  Studing X-ray spectra is a powerful diagnostic tool to infer information about  physics of stellar winds.

X-ray image  of an OB stellar association by XMM-Newton  space observatory  in the  0.2-10.0 keV energy  band. 

Stellar winds and associated mass-loss  have an extremely important influence on the evolution of massive stars, affecting evolutionary timescales, rotation, chemical profiles, surface abundances, and luminosities. A massive star settles on the main sequence as an OB-type star. When evolving, it eventually displays CNO-processed matter at its surface, and later even the products of He-burning. These two stages are spectroscopically identified with the WR-N and WR-C subtypes of Wolf-Rayet (WR) stars. A massive star ends its life with a super- or hypernova explosion. Fast-rotating metal-poor stars constitute potential gamma-ray burst progenitors (Woosley  & Heger 2006).
Winds of massive stars are driven by radiative pressure of stellar photons on the ions. This mechanism is intrinsically unstable (Lucy & White 1980). Strong shocks develop in the wind, which heat the small part of stellar wind  material to millions degrees resulting in X-ray emission (Feldmeier et al. 1997).  When two massive stars form a binary, collision of their winds may contribute to the X-ray emission. In rare magnetic stars, the confinement of stellar  winds by magnetic filed can also produce X-ray emission.
    The Orion Trapezium Cluster:  X-ray contours of the Chandra X-ray  observation overlaid onto the optical Hubble image -- from Chandra press-release  9/11/00.  Magnetic confinement of stellar wind is thought  to be the source
    of X-ray emission in the cluster most massive O star (Gagne et al. 2005)


XMM-Newton X-ray image of the sky-region 
where  WR 114 is located.



Not all stars with radiatively driven winds emit X-rays.  Rico Ignace, John C. Brown, and myself observed with XMM-Newton a WR-C type star WR114.  This stars has  strong stellar wind but was not detected  in X-rays. We further reviewed all available observations of WR-C type stars and concluded, that single stars of this type are not X-ray sources.  Later on I found that while some WR-N stars are bright X-ray sources (Ignace  et al. 2003) some other WR-N type stars, e.g.  WR61, WR157, WR16,  has not been  detected with sensitive instruments (Oskinova 2005).  Gosset et al.  (2005) reported lack of X-rays from WR-N star WR40.  Possible reason is that stellar winds in these stars are SO strong that X-rays are entirely absorbed.  However, another exciting possibility is that X-ray geneartion is different and reflects different  wind dynamics. I am currenly engaged in a work to explain puzzling differences between properties of X-ray emission among O- and WR-type stars
Detailed high resolution X-ray spectra of O-type stars were for the first time obtained by Chandra
X-ray observatory. The  Chandra ACIS-S meg  spectra of bright single non-magnetic O-type are shown on the right (Oskinova et al . 2006). These spectra have been "de-reddened",  i.e. altered in a way to look like as if there were no interstellar absorption on the line of sight towards     these stars. The spectra are dominated by emission lines and are characteristic for the optically thin plasma in collisional equilibrium heated up to few million degrees.  Meaning that all components of the gas (such as electrons and ions) have essentially the same temperature and  ions  are excited mainly by collisions with other particles.  X-ray spectra provide invaluable information about hot plasma dynamics, temperature,  density, and chemical composition. 


Only a small fraction of bolometric luminosity of O-type stars is emitted in X-rays -- 0.00001 per cent.  It means that stellar wind consists mainly from "cool" material, heated till mere 10,000 degrees.  The million degree plasma consititues only tiny fraction of general stellar wind. However, there is  interaction between X-ray photons and cool wind material.  Figure on the left shows an ultraviolet part of the spectrum of O-type supergiant  Zeta Puppis.  The black line is the spectrum obtained by IUE observatory. 
The red line is a model spectrum calculated by means of PoWR stellar atmosphere model.
The OVI  line can be reproduced by the model only when X-rays are included in the model description. 


Emission lines in spectra of massive stars are formed in fast expanding stellar wind and therefore are broad.  Black circles with error bars in Figure on the right show lines of NeX, FeVVII and OVIII   observed in the X-ray spectrum of Zeta Orionis (Alnitak).  The shape of the emission line, or line profile, is determined not only by properties of hot plasma, it is also influenced by the cool wind material, through which X-rays have to pass. It turned out that X-ray emission line profiles can be reproduced only when assuming that cool stellar wind  is clumped,  i.e. consists of density enhancements and rarifications (imagine clouds in the Earth atmosphere).  Assumption of clumped stellar wind allows to model observed X-ray lines remarkably well (red lines in the figure)  without the need to alter any of the stellar parameters.  On the other hand,  when assuming that stellar wind is smooth (like blue blue sky without any cloud) the X-ray lines cannot be modeled well (blue line).     
from  Oskinova et al. 2006