In pursuit of gamma-ray burst progenitors: the identification of a sub-population of rotating Wolf-Rayet stars

Long-duration gamma-ray bursts (GRBs) involve the most powerful cosmic explosions since the Big Bang. Whilst it has been established that GRBs are related to the death throes of massive stars, the identification of their elusive progenitors has proved challenging. Theoretical modelling suggests that rotating Wolf-Rayet (WR) stars are the best candidates. Wolf-Rayet stars are thought to be in advanced core burning stages, just prior to explosion, but their strong stellar winds shroud their surfaces, preventing a direct measurement of their rotation. Fortunately, linear spectropolarimetry may be used to probe the flattening of their winds because of stellar spin. Spectropolarimetry surveys have shown that the vast majority of WR stars (80%) have spherically symmetric winds and are therefore rotating slowly, yet a small minority (of 20%) display a spectropolarimetric signature indicative of rotation. Here we find a highly significant correlation between WR objects that carry the signature of stellar rotation and the small subset of WR stars with ejecta nebulae that have only recently transitioned from a previous red sugergiant or luminous blue variable phase. As these youthful WR stars have yet to spin-down because of mass loss, they are the best candidate GRB progenitors identified to date. When we take recently published WR ejecta nebula numbers (of Stock & Barlow 2010, MNRAS, 409, 1429), we find that five out of the six line-effect WR stars are surrounded by ejecta nebulae. The statistics imply that the null hypothesis of no correlation between line-effect WR stars and ejecta nebulae can be rejected at the 0.0004% level. Given that four line-effect and WR ejecta nebula have spectroscopically been confirmed to contain nucleo-synthetic products, we argue that the correlation is both statistically significant and physically convincing. The implication is that we have identified a sub-population of WR stars that fulfils the necessary criteria for making GRBs. Finally, we discuss the potential of identifying GRB progenitors via linear spectropolarimetry with extremely large telescopes.