Abstract
Red supergiants (RSGs) are a crucial stage in the evolution of massive stars, generally representing the last stage before a supernova explosion. During this stage, they contribute to the dynamical and chemical evolution of their host galaxy with their high continuous mass-loss rates and episodic mass-loss events. Several open questions regarding the physics of RSGs remain, such as i) the mass-loss driving mechanism and its dependency on atmospheric properties such as the metallicity and turbulence, ii) the observed lack of exploding high mass RSGs (∼ 20 − 25 M⊙ ), known as the “RSG problem”, and iii) the spectroscopic temperature discrepancies when different methodologies and diagnostics are used, extending up to 600 K when using the TiO bands as the primary temperature diagnostic. These discrepancies likely arise from limitations in the atmospheric models used to model the spectra, yet no full solutions have been presented so far (e.g., full grids of 3D RSG model atmospheres).In this ...
Red supergiants (RSGs) are a crucial stage in the evolution of massive stars, generally representing the last stage before a supernova explosion. During this stage, they contribute to the dynamical and chemical evolution of their host galaxy with their high continuous mass-loss rates and episodic mass-loss events. Several open questions regarding the physics of RSGs remain, such as i) the mass-loss driving mechanism and its dependency on atmospheric properties such as the metallicity and turbulence, ii) the observed lack of exploding high mass RSGs (∼ 20 − 25 M⊙ ), known as the “RSG problem”, and iii) the spectroscopic temperature discrepancies when different methodologies and diagnostics are used, extending up to 600 K when using the TiO bands as the primary temperature diagnostic. These discrepancies likely arise from limitations in the atmospheric models used to model the spectra, yet no full solutions have been presented so far (e.g., full grids of 3D RSG model atmospheres).In this dissertation, we explore the temperature scale for RSGs using spectroscopic observations of 137 RSGs, from which 41 show a dust-rich circumstellar environment revealed by their IR excess. This includes the currently largest set of optical RSG spectra studied beyond the Local Group (129). We obtained the surface properties of these RSGs using the MARCS atmospheric models for cool stars and compared the properties of dust-rich RSGs to dust-poor RSGs. The presence of circumstellar dust is a strong indicator of mass loss in RSGs, allowing for a connection between the inferred surface properties and mass loss. We found dusty RSGs to be more luminous, cooler, more variable, and exhibit a higher degree of self-extinction compared to RSGs without IR excess. We also explain the presence of circumstellar dust around three hot RSGs with either a potential recent mass ejection or extreme spectral variability (Levesque-Massey variables), making them candidates for future studies on extreme RSG mass loss. The molecular TiO bands, albeit shown to significantly underestimate the effective temperatures of RSGs, were the primary diagnostic used to obtain the effective temperature. To address the effective temperature discrepancy, we derived scaling relations bridging the cool TiO temperature scale to values more representative of the atmospheres of RSGs. We found that the extent of the discrepancy is tightly correlated with inferred mass-loss rates from SED fits. RSGs with higher mass-loss rates generally show a larger discrepancy between the effective temperatures obtained from the TiO band or for example spectral lines in the J −band or i −band. Although previously predicted by models, this is the first time the connection between the cool TiO bands and mass loss has been empirically established. Applying the scaling relations to the observed TiO temperatures scales the temperatures to values that agree with evolutionary predictions. We recommend further exploration of the scaling relations with the effects of metallicity and mass loss, to reliably scale temperatures of RSGs in various evolutionary stages and metallicity environments. Lastly, we conclude that the TiO diagnostic may be a promising future mass-loss probe, as current gas-dependent diagnostics (molecules in the sub-mm domain), are only resolved for nearby RSGs (up to the Magellanic Clouds).
show more