X-Shooting ULLYSES: Massive stars at low metallicity

We provide mass-loss rate predictions for O stars from Large and Small Magellanic Clouds. We calculate global (unified, hydrodynamic) model atmospheres of main sequence, giant, and supergiant stars for chemical composition corresponding to Magellanic Clouds. The models solve radiative transfer equation in comoving frame, kinetic equilibrium equations (also known as NLTE equations), and hydrodynamical equations from (quasi-)hydrostatic atmosphere to expanding stellar wind. The models allow us to predict wind density, velocity, and temperature (consequently also the terminal wind velocity and the mass-loss rate) just from basic global stellar parameters. As a result of their lower metallicity, the line radiative driving is weaker leading to lower wind mass-loss rates with respect to the Galactic stars. We provide a formula that fits the mass-loss rate predicted by our models as a function of stellar luminosity and metallicity. On average, the mass-loss rate scales with metallicity as Ṁ ~ Z0.59. The predicted mass-loss rates are lower than mass-loss rates derived from Hα diagnostics and can be reconciled with observational results assuming clumping factor Cc = 9. On the other hand, the predicted mass-loss rates either agree or are slightly higher than the mass-loss rates derived from ultraviolet wind line profiles. The calculated P v ionization fractions also agree with values derived from observations for LMC stars with Teff ≤ 40 000 K. Taken together, our theoretical predictions provide reasonable models with consistent mass-loss rate determination, which can be used for quantitative study of stars from Magellanic Clouds.

Context. Mass loss governs the evolution of massive stars and shapes the stellar surroundings. To quantify the impact of the stellar winds, we need to know the exact mass-loss rates; however, empirical constraints on the rates are hampered by limited knowledge of their small-scale wind structure, also referred to as ‘wind clumping’. Aims. We aim to improve empirical constraints on the mass loss of massive stars by investigating the clumping properties of their winds, in particular, the relation between stellar parameters and wind structure. Methods. We analysed the optical and ultraviolet spectra of 25 O-type giants and supergiants in the Large Magellanic Cloud, using the model atmosphere code FASTWIND and a genetic algorithm. We derived the stellar and wind parameters, including detailed clumping properties, such as the amount of clumping, the density of the interclump medium, velocity–porosity of the medium, and wind turbulence. Results. We obtained stellar and wind parameters for 24 of our sample stars and found that the winds are highly clumped, with an average clumping factor of 〈fcl〉 = 33 14, an interclump density factor of 〈fic〉 = 0.2 0.1, and moderate-to-strong velocity-porosity effects of 〈fvel〉 = 0.6 0.2. The scatter around the average values of the wind-structure parameters is large. With the exception of a significant, positive correlation between the interclump density factor and mass loss, we find no dependence of clumping parameters on the mass-loss rate or stellar properties. Conclusions. In the luminosity range we investigate here, the empirical and theoretical mass-loss rates both have a scatter of about 0.5 dex (or a factor 3). Within this uncertainty, the empirical rates and theoretical predictions are in agreement. The origin of the scatter of the empirically inferred mass-loss rates requires further investigation. It is possible that our description of wind clumping is still not sufficient to capture effects of the structured wind, which could contribute to the scatter.

Context. The star cluster R136 inside the Large Magellanic Cloud hosts a rich population of massive stars, including the most massive stars known. The strong stellar winds of these very luminous stars impact their evolution and the surrounding environment. We currently lack detailed knowledge of the wind structure that is needed to quantify this impact. Aims. Our goal is to observationally constrain the stellar and wind properties of the massive stars in R136, in particular the wind-structure parameters related to wind clumping. Methods. We simultaneously analyse optical and ultraviolet spectroscopy of 53 O-type and three WNh-stars using the Fastwind model atmosphere code and a genetic algorithm. The models account for optically thick clumps and effects related to porosity and velocity-porosity, as well as a non-void interclump medium. Results. We obtain stellar parameters, surface abundances, mass-loss rates, terminal velocities, and clumping characteristics and compare them to theoretical predictions and evolutionary models. The clumping properties include the density of the interclump medium and the velocity-porosity of the wind. For the first time, these characteristics are systematically measured for a wide range of effective temperatures and luminosities. Conclusions. We confirm a cluster age of 1.0–2.5 Myr and derived an initial stellar mass of ≥250 M⊙ for the most massive star in our sample, R136a1. The winds of our sample stars are highly clumped, with an average clumping factor of fcl = 29 ± 15. We find tentative trends in the wind-structure parameters as a function of the mass-loss rate, suggesting that the winds of stars with higher mass-loss rates are less clumped. We compare several theoretical predictions to the observed mass-loss rates and terminal velocities and find that none satisfactorily reproduce both quantities. The prescription of Krtička & Kubát (2018) matches the observed mass-loss rates best.

Our goal is to determine the stellar and wind properties of seven O stars in\nthe cluster NGC2244 and three O stars in the OB association MonOB2. These\nproperties give us insight into the mass loss rates of O stars, allow us to\ncheck the validity of rotational mixing in massive stars, and to better\nunderstand the effects of the ionizing flux and wind mechanical energy release\non the surrounding interstellar medium and its influence on triggered star\nformation. We collect optical and UV spectra of the target stars which are\nanalyzed by means of atmosphere models computed with the code CMFGEN. The\nspectra of binary stars are disentangled and the components are studied\nseparately. All stars have an evolutionary age less than 5 million years, with\nthe most massive stars being among the youngest. Nitrogen surface abundances\nshow no clear relation with projected rotational velocities. Binaries and\nsingle stars show the same range of enrichment. This is attributed to the youth\nand/or wide separation of the binary systems in which the components have not\n(yet) experienced strong interaction. A clear trend of larger enrichment in\nhigher luminosity objects is observed, consistent with what evolutionary models\nwith rotation predict for a population of O stars at a given age. We confirm\nthe weakness of winds in late O dwarfs. In general, mass loss rates derived\nfrom UV lines are lower than mass loss rates obtained from Ha. The UV mass loss\nrates are even lower than the single line driving limit in the latest type\ndwarfs. These issues are discussed in the context of the structure of massive\nstars winds. The evolutionary and spectroscopic masses are in agreement above\n25 Msun but the uncertainties are large. Below this threshold, the few\nlate-type O stars studied here indicate that the mass discrepancy still seems\nto hold.\n

Mass loss through stellar winds plays a dominant role in the evolution of massive stars. In particular, the mass-loss rates of very massive stars ($\gt 100\, M_{\odot}$) are highly uncertain. Such stars display Wolf–Rayet spectral morphologies (WNh), whilst on the main sequence. Metal-poor very massive stars are progenitors of gamma-ray bursts and pair instability supernovae. In this study, we extended the widely used stellar wind theory by Castor, Abbott & Klein from the optically thin (O star) to the optically thick main-sequence (WNh) wind regime. In particular, we modify the mass-loss rate formula in a way that we are able to explain the empirical mass-loss dependence on the Eddington parameter (Γe). The new mass-loss recipe is suitable for incorporation into current stellar evolution models for massive and very massive stars. It makes verifiable predictions, namely how the mass-loss rate scales with metallicity and at which Eddington parameter the transition from optically thin O star to optically thick WNh star winds occurs. In the case of the star cluster R136 in the Large Magellanic Cloud we find in the optically thin wind regime $\dot{M} \propto \Gamma _{\rm e}^{3}$, while in the optically thick wind regime $\dot{M} \propto 1/ (1 - \Gamma _{\rm e})^{3.5}$. The transition from optically thin to optically thick winds occurs at Γe, trans ≈ 0.47. The transition mass-loss rate is $\log \dot{M}~(\mathrm{M}_{\odot } \, \mathrm{yr}^{-1}) \approx -4.76 \pm 0.18$, which is in line with the prediction by Vink & Gräfener assuming a volume filling factor of $f_{\rm V} = 0.23_{-0.15}^{+0.40}$.

Context: Low- and intermediate-mass stars lose most of their stellar mass at the end of their lives on the asymptotic giant branch (AGB). Determining gas and dust mass-loss rates (MLRs) is important in quantifying the contribution of evolved stars to the enrichment of the interstellar medium. Aims: Attempt to, for the first time, spectrally resolve CO thermal line emission in a small sample of AGB stars in the Large Magellanic Cloud. Methods: ALMA was used to observe 2 OH/IR stars and 4 carbon stars in the LMC in the CO J= 2-1 line. Results: We present the first measurement of expansion velocities in extragalactic carbon stars. All four C-stars are detected and wind expansion velocities and stellar velocities are directly measured. Mass-loss rates are derived from modelling the spectral energy distribution and Spitzer/IRS spectrum with the DUSTY code. Gas-to-dust ratios are derived that make the predicted velocities agree with the observed ones. The expansion velocities and MLRs are compared to a Galactic sample of well-studied relatively low MLRs stars supplemented with "extreme" C-stars that have properties more similar to the LMC targets. Gas MLRs derived from a simple formula are significantly smaller than derived from the dust modelling, indicating an order of magnitude underestimate of the estimated CO abundance, time-variable mass loss, or that the CO intensities in LMC stars are lower than predicted by the formula derived for Galactic objects. This could be related to a stronger interstellar radiation field in the LMC. Conclusions: Although the LMC sample is small and the comparison to Galactic stars is non-trivial because of uncertainties in their distances it appears that for C stars the wind expansion velocities in the LMC are lower than in the solar neighbourhood, while the MLRs appear similar. This is in agreement with dynamical dust-driven wind models.

We calculate NLTE line-driven wind models of selected O stars in the spectral range of O4 to O9 in the Small Magellanic Cloud (SMC). We compare predicted basic wind properties, i.e. the terminal velocity and the mass-loss rate with values derived from observation. We found relatively good agreement between theoretical and observed terminal velocities. On the other hand, predicted mass-loss rates and mass-loss rates derived from observation are in a good agreement only for higher mass-loss rates. Theoretical mass-loss rates lower than approximately 10^{-7} M_o/year are significantly higher than those derived from observation. These results confirm the previously reported problem of weak winds, since our calculated mass-loss rates are in a fair agreement with predictions of Vink et al. (2001). We study multicomponent models for these winds. For this purpose we develop a more detailed description of wind decoupling. We show that the instability connected with the decoupling of individual wind elements may occur for low-density winds. In the case of winds with very low observed mass-loss rate the multicomponent effects are important for the wind structure, however this is not able to consistently explain the difference between predicted mass-loss rate and mass-loss rate derived from observation for these stars. Similar to previous studies, we found the dependence of wind parameters on the metallicity. We conclude that the wind mass-loss rate significantly increases with metallicity as dM/dt ~ Z^0.67, whereas the wind terminal velocity on average depends on metallicity only slightly, namely v_infty ~ Z^0.06 (for studied stars).

Our limited understanding of the winds of massive stars hampers our ability to predict their evolution. We currently often rely on mass-loss rate prescriptions that show strong features that lack a firm empirical confirmation, such as the bi-stability jump. This bi-stability jump is a pronounced increase in mass loss in the decreasing temperature regime T_ ̊m eff ∼28-21,kK, i.e., spectral types B1-B2. Different prescriptions also give different metallicity dependences; these sometimes include a dependence on effective temperature, luminosity or both, that may also differ among prescriptions. Although recent papers have compared empirical results to prescriptions, a large observational sample of B stars for which the wind has been scrutinised over different metallicity ranges is still lacking. Through the modelling of both ultraviolet (from the ULLYSES programme) and optical (from the XShootU programme) high-resolution spectra, we determined the stellar and wind parameters, including detailed clumping parameters, of 24 SMC B stars ranging in T_ eff from 13 to 29,kK. By combining this sample with LMC studies, we compared the wind behaviour of B stars in two different metallicity regimes. We compared our results to common mass-loss rate prescriptions to test features present in these models and their metallicity dependence. We have used the model atmosphere code fastwind and the genetic algorithm code Kiwi-GA to fit the UV and optical spectra simultaneously. This allows us to determine wind properties including clumping behaviour. The moderate to strong metallicity trends present in the mass-loss prescriptions (dot M ∝ Z^ 0.41-1.4 ) explored here overestimate the empirical metallicity dependence in the B-star regime, which appears very weak. Similarly to a previous study of B supergiants in the LMC, we do not find a sudden increase in mass-loss rate at approximately spectral type B1, but only a weak temperature trend. As in this LMC study, we show that on average around 40% of the wind mass is located in the wind medium between the clumps. Using other studies from this paper series, we compiled a sample of 80+ O and B stars in the SMC and LMC. From a comparison we find a clear difference in O- and B-type metallicity dependence. The apparent lack of a bi-stability jump in the B-star regime and a very weak metallicity dependence for the same stars offers new empirical constraints to theoretical models of line-driven winds. As differences between these models are large (reaching orders of magnitude in parts of parameter space) such constraints are much needed. More studies exploring the mass-loss rates of B stars and cooler objects will be helpful to our understanding of hot star winds.

We employ newly computed grids of spectra reprocessed by dust for estimating the total dust production rate (DPR) of carbon stars in the Small Magellanic Cloud (SMC). For the first time, the grids of spectra are computed as a function of the main stellar parameters, i.e. mass-loss rate, luminosity, effective temperature, current stellar mass and element abundances at the photosphere, following a consistent, physically grounded scheme of dust growth coupled with stationary wind outflow. The model accounts for the dust growth of various dust species formed in the circumstellar envelopes of carbon stars, such as carbon dust, silicon carbide and metallic iron. In particular, we employ some selected combinations of optical constants and grain sizes for carbon dust that have been shown to reproduce simultaneously the most relevant colour–colour diagrams in the SMC. By employing our grids of models, we fit the spectral energy distributions of ≈3100 carbon stars in the SMC, consistently deriving some important dust and stellar properties, i.e. luminosities, mass-loss rates, gas-to-dust ratios, expansion velocities and dust chemistry. We discuss these properties and we compare some of them with observations in the Galaxy and Large Magellanic Cloud. We compute the DPR of carbon stars in the SMC, finding that the estimates provided by our method can be significantly different, between a factor of ≈2–5, than the ones available in the literature. Our grids of models, including the spectra and other relevant dust and stellar quantities, are publicly available at http://starkey.astro.unipd.it/web/guest/dustymodels.

We use the Auriga cosmological simulations of Milky Way (MW)-mass galaxies and their surroundings to study the satellite populations of dwarf galaxies in ΛCDM. As expected from prior work, the number of satellites above a fixed stellar mass is a strong function of the mass of the primary dwarf. For galaxies as luminous as the Large Magellanic Cloud (LMC), and for halos as massive as expected for the LMC (from its rotation speed), the simulations predict about ∼3 satellites with stellar masses exceeding M* > 105 M⊙. If the LMC is on its first pericentric passage, then these satellites should be near the LMC and should have orbital angular momenta roughly coincident with that of the LMC. We use 3D positions and velocities from the 2nd data release of the Gaia mission to revisit which of the “classical” MW dwarf spheroidals could plausibly be LMC satellites. The new proper motions of the Fornax and Carina dwarf spheroidals place them on orbits closely aligned with the orbital plane of the Magellanic Clouds, hinting at a potential Magellanic association. Together with the Small Magellanic Cloud (SMC), this result raises to 3 the number of LMC satellites with M* > 105 M⊙, as expected from simulations. This also fills the 12-mag luminosity gap between the SMC and the ultra-faints Hyi1, Car2, Hor1, and Car3, the few ultra-faint satellites confirmed to have orbits consistent with a Magellanic origin.

It is yet well understood how mass-loss rates from evolved stars depend on metallicities. With a half of the solar metallicity and the distance of only 50 kpc, the evolved stars of the Large Magellanic Cloud (LMC) are an ideal target for studying mass loss at low metallicity. We have obtained spectra of red-supergiants in the LMC, using the Hershel Space Observatory, detecting CO thermal lines fro J=6–5 up to 15–14 lines. Modelling CO lines with non-LTE Radiative transfer code suggests that CO lines intensities can be well explained with high gas-to-dust ratio, with no obvious reduction in mass-loss rate at the LMC. We conclude that the luminosities of the stars are primary factors on mass-loss rates, rather than the metallicity.

A nearly complete sample of 24 Magellanic Cloud WC/WO subclass Wolf–Rayet stars is studied spectroscopically and photometrically to determine its binary frequency. Theory predicts the Roche lobe overflow produced Wolf–Rayet binary frequency to be 52±14 per cent in the Large Magellanic Cloud and 100 per cent in the Small Magellanic Cloud, not counting non-Roche lobe overflow Wolf–Rayet binaries. Lower ambient metallicity (Z) leads to lower opacity, preventing all but the most massive (hence luminous) single stars from reaching the Wolf–Rayet stage. However, theory predicts that Roche lobe overflow even in binaries of modest mass will lead to Wolf–Rayet stars in binaries with periods below approximately 200 d, for initial periods below approximately 1000 d, independent of Z. By examining their absolute continuum magnitudes, radial velocity variations, emission-line equivalent widths and full widths at half-maximum, a WC/WO binary frequency of only 13 per cent, significantly lower than the prediction, is found in the Large Magellanic Cloud. In the unlikely event that all of the cases with a less certain binary status actually turn out to be binary, current theory and observation would agree. (The Small Magellanic Cloud contains only one WC/WO star, which happens to be a binary.) The three WC+O binaries in the Large Magellanic Cloud all have periods well below 1000 d. The large majority of WC/WO stars in such environments apparently can form without the aid of a binary companion. Current evolutionary scenarios appear to have difficulty explaining either the relatively large number of Wolf–Rayet stars in the Magellanic Clouds, or the formation of Wolf–Rayet stars in general.

Accurate mass-loss rates and terminal velocities from massive stars winds are essential to obtain synthetic spectra from radiative transfer calculations and to determine the evolutionary path of massive stars. From a theoretical point of view, analytical expressions for the wind parameters and velocity profile would have many advantages over numerical calculations that solve the complex non-linear set of hydrodynamic equations. In a previous work, we obtained an analytical description for the fast wind regime. Now, we propose an approximate expression for the line-force in terms of new parameters and obtain a velocity profile closed-form solution (in terms of the Lambert W function) for the δ-slow regime. Using this analytical velocity profile, we were able to obtain the mass-loss rates based on the m-CAK theory. Moreover, we established a relation between this new set of line-force parameters with the known stellar and m-CAK line-force parameters. To this purpose, we calculated a grid of numerical hydrodynamical models and performed a multivariate multiple regression. The numerical and our descriptions lead to good agreement between their values.

Context. Massive stars emit copious amounts of radiation, profoundly affecting their environment in galaxies and contributing to the reionization of the Universe. However, their evolution and thus their ionizing feedback are still not fully understood. One of the largest gaps in current stellar evolution calculations is the lack of a model for the mass ejections that occur when the stars reach the Eddington limit, such as during a Luminous Blue Variable (LBV) phase. Aims. Here, we aim to remedy this situation by providing a physically motivated and empirically calibrated method applicable in any 1D stellar evolution code to approximate the effect of such mass loss on stellar evolution. Methods. We employed the 1D stellar evolution code MESA, in which we implement a new mass-loss prescription that becomes active when stellar models inflate too much when reaching the Eddington limit. We used lines of constant inflation factors in the Hertzsprung-Russell diagram (HRD) for a simple empirical calibration of the threshold value. We calculated synthetic massive-star stellar populations using grids of single-star models with this mass loss prescription compared them with the observed populations in the Large and Small Magellanic Clouds. Further, with already computed grids of binary evolution models, we investigated the impact of binarity on our predictions. Results. Our single-star models reproduce key features of the observed stellar populations, namely, (i) the absence of stars located beyond the Humphreys-Davidson limit; (ii) an upper limit of red supergiant (RSG) luminosities; (iii) the faintest observed single Wolf-Rayet (WR) stars; (iv) the absolute number of O-stars, WRs, and RSGs; (v) WO stars in low metallicity environments; and (vi) the positions of LBV stars in the HRD. We show that binarity still plays an important role in explaining the observed WR stars. However, a large fraction of the binary population can also be explained via self-stripping. At the same time, our binary population explains the 70% binary fraction of O-stars and the 40% binary fraction of WR stars. However, our synthetic population also has caveats, such as an overproduction of bright H-free WN stars. Conclusions. Our results show that the effect of the Eddington-limit induced mass ejections on the structure and evolution of massive stars can remove the tension between predicted and observed massive star populations. A more fundamental treatment of these effects, particularly for hydrogen-poor stars, is needed to fully comprehend massive star evolution.

Hot, massive stars are known to host unstable, radiation-driven outflowing winds, giving rise to dense clumps of material which severely affect the diagnostic techniques used to derive wind properties of massive stars. Most of the current diagnostic models account for wind inhomogeneities by assuming a one-component medium consisting of optically thin clumps, and maintaining a smooth velocity-field. However, this neglects important light-leakage effects through porous channels in-between the clumps. These light-leakage effects have recently been incorporated in the stellar atmosphere modelling code FASTWIND, and here we will present quantitative mass-loss results from a combined Ultraviolet-Optical wind analysis of O-supergiants in the Galaxy. Using a genetic-algorithm fitting-approach, we systematically investigate the impact the wind physics has on derived stellar and wind parameters, and how this depends on metallicity and spectral type. We compare our findings with earlier results (which do not take into account such light-leakage effects), to standard mass-loss rates usually included in evolution model studies of massive stars, and with theoretical predictions of clumping properties. We will also present the first systematic empirical constraints on the new wind parameters, associated with light-leakage, and compare these with theoretical predictions.