Wolf‐Rayet Phenomena in Massive Stars and Starburst Galaxies: IAU Symposium 193¹

The international hot star community organized IAU Symposium 193 at the Camino Real Hotel in Puerto Vallarta as part of their continuing series of "beach symposia." There was very active participation from astronomers from Mexico, who hosted the conference; 186 individuals attended from 24 countries, and 54 oral papers and 130 posters were presented. The weather was warm and sunny, the surroundings peaceful, and informal discussions occurred while dining, drinking, and sitting on the beach. The poster sessions colocated with the meeting rooms were particularly well attended and remained up for the entire conference. This summary is a personal appraisal, one which might well be very different in the hands of another reporter.

W‐R phenomena as strong broad emission lines were first identified in stars in 1867. We now know that this is due to intense stellar winds. W‐R stars are initially high‐mass, very luminous objects that are undergoing processes that result in a great deal of mass loss, much of which is currently observable. Models for single‐star evolution with extensive mass loss indicate that the anomalous composition of W‐R stars (subtypes: WN, nitrogen‐ and helium‐rich; WC, helium‐ and carbon‐rich) is a result of core burning products being present at the stellar "surface" owing to previous mass loss and/or mixing. Starburst galaxies are those systems in which strong (narrow) nebular recombination lines are seen, indicating the presence of a large number of hot exciting stars, mostly of type O. Those starburst galaxies with broad emission lines in the integrated spectrum from W‐R stars are referred to as W‐R galaxies. Roughly 70% of the Symposium was devoted to issues concerned with W‐R stars; the remainder, to that of galaxies. It is a measure of the explosion of studies in the past few years in this field that this was by far the best attended "hot star" Symposium and one in which there was a substantial interaction between astrophysicists with "stellar" and "galaxy" backgrounds.

The imminent publication of new catalogs of W‐R stars in the Galaxy and Magellanic Clouds was announced; the census amounts to 211 stars in the former, 132 in the LMC, and nine in the SMC. An analysis of the W‐R populations of the other Local Group galaxies was given. A new catalog of W‐R galaxies lists 126 members, indicating a doubling time of 3.5 yr since the initial serendipitous discovery in 1976. W‐R stars radiate from the X‐ray to radio region, which is, from the UV longward, mostly thermal, but anomalous strong nonthermal emission is found in some cases. The thermal radiation is primarily emitted from dramatically different levels in the strong stellar winds; a "photosphere" is not present. While most objects appear to have spherically symmetric wind geometry, a small fraction of the objects are asymmetric; high angular momentum may play a role. Most (perhaps all) of the stellar winds are nonhomogeneous in which "clumping" is present (providing the X‐ray emission from the resultant shocks) and lowering the inferred mass‐loss rates compared to that obtained with an assumption of uniform flow. Many stars also show variability at low levels. It is fair to say that the physics of these variable phenomena is not yet well understood, but considerable progress was reported.

The driving mechanism of the winds of luminous OB stars is well established to be radiative in nature due to the myriad of lines present in the bright UV part of the spectrum. A "single‐scattering" mode of photons is sufficient to account for the winds of these stars, but the inferred momentum of W‐R star winds had seemed to have been too large given their luminosities and mass‐loss rates. Several lines of evidence at the conference suggested that the W‐R star luminosities were higher, and the mass‐loss rates lower, than previously inferred. Up to 10 multiple scatterings would be required for radiatively driven winds, which seems within the realm of possibilities in the energetics. Detailed predictions are needed.

The physics of W‐R star winds is becoming clearer with improved observations of line profiles and comparison to theoretical predictions. The "standard model" of non‐LTE detailed line transfer with H and He composition (only) has been newly enhanced with the addition of metal (Z) line blanketing, in which substantial effects (e.g., "backwarming") are predicted, and observed, in the emergent continuum radiation. In particular, the emission below the He ii ionization edge is substantially affected and highly dependent on Z. In low‐Z objects, sufficient radiation can be emitted to doubly ionize helium in the surrounding H ii region, thus explaining the presence of a narrow He ii λ4686 recombination line. Some low‐Z W‐R galaxies show this feature in addition to the normal recombination lines of hydrogen and (neutral) helium.

Single‐star evolution models of W‐R stars generated by several groups have been able to predict many of their observational aspects. The importance of "mixing" of material from the core to the "surface" has recently become appreciated, and new results were presented. Both the mass loss and differential rotation (meridional circulation leading to mixing) in W‐R stars and their predecessors are important. While the details are not yet entirely in place, the overall picture is becoming clearer.

Binary W‐R systems provide both opportunities and problems for stellar astrophysicists. Orbital solutions lead to estimates of the masses of the stars. It was realized long ago that W‐R stars are typically overluminous for their masses, thus establishing the concept of their helium burning nature. There is an extensive literature of orbits of Galactic W‐R stars; those in the Magellanic Clouds are just now being produced. Close binary evolution undoubtedly will influence how a massive star would have evolved compared to what would have been were it single. The close binary fraction and its overall impact on the evolution of massive stars remains, however, controversial. Wide binary W‐R stars provide an opportunity to explain the nonthermal emission as originating in the colliding winds of the two hot components. While the physics is not totally yet understood, considerable progress in observations and in the theory was presented. Close binary interaction effects can have major impact on observations over all wavelength bands. The very luminous object HD 5980 in the SMC is an example of this effect. Indeed, a separate informal workshop on this luminous blue variable (LBV) plus WN star was scheduled the day before the Symposium. This short‐period eccentric orbit binary, which rivals η Car in luminosity, recently went through a substantial outburst, coupled with a change in the overall spectral type. It is fair to say that while the event is somehow related to the binary interaction, the system is far from being understood.

The substantial Lyman continuum emission and the strong stellar winds of W‐R stars affect their environments, often with observational consequences. The ionized Strömgren spheres may be analyzed for, say, the composition. The stellar winds may overtake and excite previously slower moving ejected material. So‐called ring nebulae are observed surrounding some W‐R stars. In many cases, the composition and outward velocity of the ring material can be identified with a previous red supergiant phase for the W‐R star. Probably the initially most luminous stars do not go through such a phase but rather lose mass both while an O star and in a subsequent not well modeled LBV process.

Luminous clusters of hot stars found within the Local Group are the bridge between nearby individually resolved objects and the distant, more luminous starburst galaxies. Observations of NGC 3603 and some clusters in and near the center of the Galaxy, R136‐30 Dor in the LMC and NGC 604 in M33, were presented. Each has a number of very luminous O and W‐R type stars. Solar mass stars on a tight sequence are identified in NGC 3603, which shows that low‐mass objects can be produced in the presence of very luminous hot stars. These clusters have a few hundred O stars each, sometimes referred to as "mini‐starbursts." They are at the low‐luminosity end of starbursts.

The parameters of starburst galaxies, that is their ages, burst durations, stellar populations, etc., may be directly inferred from observations of the integrated emergent continua and their stellar spectral line features. Analyses of the nebular recombination lines provide indirect estimates of the numbers of exciting (OB+W‐R) stars. Spectral synthesis models have been provided by several groups to predict these various signatures. The stars and ionized gas may not be cospatial in some cases; thus caution needs to be exercised when comparing the various methods, although for a number of objects studied the interagreement is not bad. From the predictions of the single‐star evolutionary models, one would expect the presence of W‐R stars in starbursts with ages from between 3 and 6 Myr, in rough agreement with the observations. The frequency of WN and WC types also follows expectations; future work will need to include the effects of stellar mixing. The impact of close binary evolution on starburst parameters might be important if this fraction is large.

In many (perhaps all) examples of starbursts examined with high spatial resolution, the luminosity comes from smaller but even more intense starburst knots; these have been called "super–star clusters" (SSCs). These may evolve into globular clusters as had been suggested in the literature, but this is not accepted by all. Hubble Space Telescope observations and analyses of SSCs in several starburst galaxies were presented. Most of these objects seem to be relatively young, not unexpected considering the sample was of W‐R galaxies. The SSC range in luminosity (and mass) upward from R136 by a factor of 100. Their colors and luminosity functions are similar to those SSCs observed in other well‐known merger galaxies once account is taken of their youth.

W‐R starburst galaxies may all be the result of recent mergers and interactions, as several examples of objects previously thought to be isolated were shown to have fainter nearby companions. It is suggestive that this interaction feature has also been invoked to explain the luminous events in IR‐luminous galaxies, and there may be a generic relationship. There is a clear similarity in spectral morphology between UV spectra of relatively nearby SSCs in (W‐R) starburst galaxies, and those star‐forming regions recently found at high redshift (z). W‐R phenomena, as indicative of intense star formation activity, are also noted in connection with AGNs. The presence of such events near to an active nucleus is surprising, to say the least.

Not only are W‐R phenomena in massive stars a natural consequence of stellar evolution, but a similar appearance in starburst galaxies may always follow "if the time is right." Studying nearby stars and individually resolved luminous clusters can shed light on more energetic, but more distant, starburst galaxies, and conversely.