Context. Massive stars play an important role in both cluster and galactic evolution and the rate at which they lose mass is a key driver of both their own evolution and their interaction with the environment up to and including their terminal SNe explosions. Young massive clusters provide an ideal opportunity to study a co-eval population of massive stars, where both their individual properties and the interaction with their environment can be studied in detail. Aims. We aim to study the constituent stars of the Galactic cluster Westerlund 1 in order to determine mass-loss rates for the diverse post-main sequence population of massive stars. Methods. To accomplish this we made 3mm continuum observations with the Atacama Large Millimetre/submillimetre Array. Results. We detected emission from 50 stars in Westerlund 1, comprising all 21 Wolf-Rayets within the field of view, plus eight cool and 21 OB super-/hypergiants. Emission nebulae were associated with a number of the cool hypergiants while, unexpectedly, a number of hot stars also appear spatially resolved. Conclusions. We were able to measure the mass-loss rates for a unique population of massive post-main sequence stars at every stage of evolution, confirming a significant increase as stars transitioned from OB supergiant to WR states via LBV and/or cool hypergiant phases. Fortuitously, the range of spectral types exhibited by the OB supergiants provides a critical test of radiativelydriven wind theory and in particular the reality of the bi-stability jump. The extreme mass-loss rate inferred for the interacting binary Wd1-9 in comparison to other cluster members confirmed the key role binarity plays in massive stellar evolution. The presence of compact nebulae around a number of OB and WR stars is unexpected; by analogy to the cool super-/hypergiants we attribute this to confinement and sculpting of the stellar wind via interaction with the intra-cluster medium/wind. Given the morphology of core collapse SNe depend on the nature of the pre-explosion circumstellar environment, if this hypothesis is correct then the properties of the explosion depend not just on the progenitor, but also the environment in which it is located.

Context. Massive stars and their stellar winds are important for a number of feedback processes. The mass lost in the stellar wind can help determine the end-point of the star as a NS or a BH. However, the impact of mass-loss on the post-Main Sequence evolutionary stage of massive stars is not well understood. Westerlund 1 is an ideal astrophysical laboratory in which to study massive stars and their winds in great detail over a large range of different evolutionary phases. Aims. We aim to study the radio emission from Westerlund 1, in order to measure radio fluxes from the population of massive stars, and determine mass-loss rates and spectral indices where possible. Methods. Observations were carried out in 2015 and 2016 with the Australia telescope compact array (ATCA) at 5.5 and 9 GHz using multiple configurations, with maximum baselines ranging from 750 m to 6 km. Results. 30 stars were detected in the radio from the fully concatenated dataset, 10 of which were WRs (predominantly late type WN stars), 5 YHGs, 4 RSGs, 1 LBV star, the sgB[e] star W9, and several O and B supergiants. New source detections in the radio were found for 5 WR stars, and 5 OB supergiants. These detections have led to evidence for 3 new OB supergiant binary candidates, inferred from derived spectral index limits. Conclusions. Spectral indices and index limits were determined for massive stars in Westerlund 1. For cluster members found to have partially optically thick emission, mass-loss rates were calculated. Under the approximation of a thermally emitting stellar wind and a steady mass-loss rate, clumping ratios were then estimated for 8 WRs. Diffuse radio emission was detected throughout the cluster. Detections of knots of radio emission with no known stellar counterparts indicate the highly clumped structure of this intra-cluster medium, likely shaped by a dense cluster wind.

We report new 5.5 GHz radio observations of the massive star cluster Westerlund 1, taken by the Australia Telescope Compact Array, detecting nine of the ten yellow hypergiants (YHGs) and red supergiants (RSGs) within the cluster. Eight of nine sources are spatially resolved. The nebulae associated with the YHGs Wd1-4a, -12a and -265 demonstrate a cometary morphology -the first time this phenomenon has been observed for such stars. This structure is also echoed in the ejecta of the RSGs Wd1-20 and -26; in each case the cometary tails are directed away from the cluster core. The nebular emission around the RSG Wd1-237 is less collimated than these systems but once again appears more prominent in the hemisphere facing the cluster. Considered as a whole, the nebular morphologies provide compelling evidence for sculpting via a physical agent associated with Westerlund 1, such as a cluster wind.

Despite their large impact on stellar and galactic evolution, the properties of outflows from red supergiants are not well characterised. We used the Onsala 20m telescope to perform a spectral survey at 3 mm and 4 mm (68-116  GHz) of the red supergiant NML Cyg, alongside the yellow hypergiant IRC +10420. Our observations of NML Cyg were combined with complementary archival data to enable a search for signatures of morphological complexity in the circumstellar environment, using emission lines from 15 molecular species. The recovered parameters imply the presence of three distinct, coherent and persistent components, comprised of blue-shifted and red-shifted components, in addition to an underlying outflow centred at the stellar systemic velocity. Furthermore, to reproduce 12CO emission with three-dimensional radiative transfer models required a spherical outflow with three superposed conical outflows, one towards and one away from the observer, and one in the plane of the sky. These components are higher in density than the spherical outflow by up to an order of magnitude. We hence propose that NML Cyg’s circumstellar environment consists of a small number of high-density large-scale coherent outflows embedded in a spherical wind. This would make the mass-loss history similar to that of VY CMa, and distinct from μ Cep, where the outflow contains many randomly distributed smaller clumps. A possible correlation between stellar properties, outflow structures and content is critical in understanding the evolution of massive stars and their environmental impact.