The Universe's Prolific Recyclers

Massive stars (those with masses greater than eight times that of our Sun) live short, violent lives, with lifetimes of only a few million years (compare this to our Sun which is currently 5 billion years old!). During their brief existence, they contribute profoundly to the chemical and dynamical makeup of galaxies via their strong, metal-enriched stellar winds in life and their explosive supernovae outbursts in death. They churn out much of the material that was condensed to make them, in effect recycling to make way for the next generation of stars. They are essential in creating the heavier elements that constitute both planets and our very selves. The effect massive stars have on their surrounding is nicely illustrated in Figure 1 below showing the star-forming galaxy NGC4214.

Figure 1: HST  Hubble Heritage image of the starburst galaxy NGC4214 showing the impact of massive stars on their surroundings. Each cloud in this image is illuminated due to the strong ultraviolet light emitted by the hot, ionizing stars embedded within them. The powerful stellar winds ejected by these stars give rise to `wind blown bubbles' (or cavities) as the winds plough their way through the ISM, possibly triggering new star formation.

Our understanding of these powerhouses is key to our knowledge of galactic evolution. Despite many decades of effort, we find that the winds emanating from massive stars remain somewhat enigmatic. As the long-held assumption of a smooth, homogeneous outflow continues to crumble, work at UCL and in other international research groups is attempting to explain and quantify the ‘clumpy’ winds of these cosmic recyclers, and how they affect the wider Universe.


One of the key parameters defining the lifetime and impact of a massive star is its rate of mass-loss. Fused carbon, nitrogen, oxygen and other elements are ejected at around 10-7 solar masses per year (around 1021 kg a day), according to canonical hydrodynamic calculations (Vink et al. 2001). This occurs due to considerable momentum injected by the intense radiation field present. During the last several years, it has become clear that due to intrinsic instabilities in the driving of mass from the surface of a star, and to shocks forming in the wind, there is wind structure that is disrupting our empirical measurements. At UCL we have adapted a mass-loss measuring technique, which highlights how grossly we may have been over-estimating mass-loss, and have begun to theorise what the real situation might be.

The Carbon Project

Sophisticated computer codes are used to model the atmospheres and winds of massive stars. They solve the equations which govern the system, predict how the intensity spectrum might look, and how much of each ion is present at different depths in the atmosphere. We used one such code, CMFGEN (Hillier et al. 1998), to deduce the fraction of C3+ in the wind at different temperatures. This ion was found to account for more than half of all carbon for stars with temperatures between 32500 and 37000 K. A second code was used to fit real stellar carbon spectra empirically, and predict the product of mass-loss rate and ion fraction. These two quantities cannot usually be disentangled, but since C3+ was known to be dominant the ion fraction could be assumed to be unity. This gave us the rate of mass-loss for around 30 of the hottest massive stars. They were all in the range 10-9–10-10 solar masses per year – up to three orders of magnitude lower than the canonical values. Are massive stars recycling rather less than we thought?

CIV Wind Line Profile
Figure 2: This plot shows the empirical fit (red) to observed ultraviolet spectral data (white). Wavelength is in Angstroms.

Clumping and Porosity

Some diagnostics of mass-loss appear to concur with the hydrodynamic predictions, whilst others seem to derive quite low rates. Including a formulation for a ‘clumped’ wind in the models has led to a downward revision of those higher rates, whilst the concept of ‘porosity’ deals with much lower rates.  Porosity assumes that a stellar wind is composed of large, optically thick clumps, and that some electromagnetic flux can penetrate through the gaps whilst some is stopped short by the clump material. Modelling of this by Oskinova et al. (2007) shows that neglecting the effect of porosity leads to underestimating mass-loss by a factor of a few. This is insufficient to account for the differences seen in the Carbon Project however. New work is clearly required in the way models treat stellar winds, if this problem is to be reconciled.


Massive stars continue to be viewed as integral elements in the make-up and ongoing evolution of galaxies, but the quantification of their interaction with their environment is unreconciled. Projects like this one highlight how we are currently unable to model massive stellar winds satisfactorily, to the point where different outflow diagnostics do not agree within uncertainties. Further endeavour is hence needed to bridge the gap between theory and observation, and it will undoubtedly involve a more thorough treatment of clumping.

Stars like this may be some of the oldest subjects of study within astronomy, but they continue to throw up interesting and challenging problems. It may well be many years before we can claim a comprehensive understanding of these fascinating objects.

Thanks to Matt Austin for this article.

Back to Top

Page last modified on 16 jul 10 14:39 by Fabrizio Sidoli