Eta Carinae is the most massive active binary within 10,000 light-years and is famous for the largest non-terminal stellar explosion ever recorded. Observations reveal that the supermassive (~120 M⊙) binary, consisting of an LBV and either a WR or extreme O star, undergoes dramatic changes every 5.54 years due to the stars’ very eccentric orbits (e ≈ 0.9). Many of these changes are caused by a dynamic wind-wind collision region (WWCR) between the stars, plus expanding fossil WWCRs formed one, two, and three 5.54-year cycles ago. The fossil WWCRs can be spatially and spectrally resolved by the Hubble Space Telescope/Space Telescope Imaging Spectrograph (HST/STIS). Starting in June 2009, we used the HST/STIS to spatially map Eta Carinae’s fossil WWCRs across one full orbit, following temporal changes in several forbidden emission lines (e.g. [Feiii] 4659 Å, [Feii] 4815 Å), creating detailed data cubes at multiple epochs. Multiple wind structures were imaged, revealing details about the binary’s orbital motion, photoionization properties, and recent (~5 − 15 year) mass-loss history. These observations allow us to test 3-D hydrodynamical and radiative-transfer models of the interacting winds. Our observations and models strongly suggest that the wind and photoionization properties of Eta Carinae’s binary have not changed substantially over the past several orbital cycles. They also provide a baseline for following future changes in Eta Carinae, essential for understanding the late-stage evolution of this nearby supernova progenitor. For more details, see Gull et al. (2016) and references therein.

The evolution of massive stars encompasses short-lived transition phases in which mass-loss is more enhanced and usually eruptive. A complex environment, combining atomic, molecular and dust regions, is formed around these stars. In particular, the circumstellar environment of B[e] Supergiants is not well understood. To address that, we have initiated a campaign to investigate their environments for a sample of Galactic and Magellanic Cloud sources. Using high-resolution optical and near-infrared spectra (MPG-ESO/FEROS, GEMINI/Phoenix and VLT/CRIRES, respectively), we examine a set of emission features ([OI], [CaII], CO bandheads) to trace the physical conditions and kinematics in their formation regions. We find that the B[e] Supergiants are surrounded by a series of rings of different temperatures and densities, a probable result of previous mass-loss events. In many cases the CO forms very close to the star, while we notice also an alternate mixing of densities and temperatures (which give rise to the different emission features) along the equatorial plane.

The WR stellar population can be distinguished, at least partially, from other stellar populations by broad-band IR colour selection. We present the use of a machine learning classifier to quantitatively improve the selection of Galactic Wolf-Rayet (WR) candidates. These methods are used to separate the other stellar populations which have similar IR colours. We show the results of the classifications obtained by using the 2MASS J, H and K photometric bands, and the Spitzer/IRAC bands at 3.6, 4.5, 5.8 and 8.0μm. The k-Nearest Neighbour method has been used to select Galactic WR candidates for observational follow-up. A few candidates have been spectroscopically observed. Preliminary observations suggest that a detection rate of 50% can easily be achieved.

The role of episodic mass loss in evolved massive stars is one of the outstanding questions in stellar evolution theory. Integral field spectroscopy of nebulae around massive stars provide information on their recent mass-loss history. η Car is one of the most massive evolved stars and is surrounded by a complex circumstellar environment. We have conducted a three-dimensional morpho-kinematic analysis of η Car’s ejecta outside its famous Homunculus nebula. SHAPE modelling of VLT MUSE data establish unequivocally the spatial cohesion of the outer ejecta and the correlation of ejecta with the soft X-ray emission.

We present the first successful simulations of neutrino-driven supernova explosions in three dimensions (3D) using the Vertex-Prometheus code including sophisticated energy-dependent neutrino transport. The simulated models of 9.6 and 20 solar-mass iron-core stars demonstrate that successful explosions can be obtained in self-consistent 3D simulations, where previous models have failed. New insights into the supernova mechanism can be gained from these explosions. The first 3D model (Melson et al. 2015a) explodes at the same time but more energetically than its axially symmetric (2D) counterpart. Turbulent energy cascading reduces the kinetic energy dissipation in the cooling layer and therefore suppresses neutrino cooling. The consequent inward shift of the gain radius increases the gain layer mass, whose recombination energy provides the surplus for the explosion energy.

The second explosion (Melson et al. 2015b) is obtained through a moderate reduction of the neutral-current neutrino opacity motivated by strange-quark contributions to the nucleon spin. A corresponding reference model without these corrections failed, which demonstrates how close current 3D models are to explosion. The strangeness adjustment is meant as a prototype for remaining neutrino opacity uncertainties.

We carried out a pivot experiment to select distant luminous late-type stars on the basis on their 2MASS and GLIMPSE photometry. Low-resolution infrared spectra enabled us to measure the equivalent widths (EWs) of their CO band-heads at 2.293 μm, and to confirm an extraordinarily high detection rate of red supergiants (RSGs), i.e. 61% (Messineo et al. (2016)).

We discuss light-curve and spectral properties of ultra-stripped core-collapse supernovae. Ultra-stripped supernovae are supernovae with ejecta masses of only ~0.1M⊙ whose progenitors lose their envelopes due to binary interactions with their compact companion stars. We follow the evolution of an ultra-stripped supernova progenitor until core collapse and perform explosive nucleosynthesis calculations. We then synthesize light curves and spectra of ultra-stripped supernovae based on the nucleosynthesis results. We show that ultra-stripped supernovae synthesize ~0.01M⊙ of the radioactive 56Ni, and their typical peak luminosity is around 1042 erg s−1 or −16 mag. Their typical rise time is 5 − 10 days. By comparing synthesized and observed spectra, we find that SN 2005ek and some of so-called calcium-rich gap transients like PTF10iuv may be related to ultra-stripped supernovae.

The objective is to determine the nature of the unseen companion of the single-lined spectroscopic binary, WR 148 (= WN7h+?). The absence of companion lines supports a compact companion (cc) scenario. The lack of hard X-rays favours a non-compact companion scenario. Is WR 148 a commonplace WR+OB binary or a rare WR+cc binary?

The next Galactic supernova is expected to bring great opportunities for the direct detection of gravitational waves, full flavor neutrinos, and multi-wavelength photons. To prepare for appropriate observations of these multi-messenger signals, we use a long-term numerical simulation of the core-collapse supernova and discuss detectability of the signals in different situations. By exploring the sequential multi-messenger signals of a nearby CCSN, we discuss preparations for maximizing successful studies of such an unprecedented stirring event.

In 2006, the Of?p star HD191612 became the second O-star where a magnetic field was discovered. It provided a benchmark to understand the Of?p phenomenon as a whole. Ten years later, an X-ray monitoring performed at high-resolution reveals the behaviour of the hottest magnetospheric plasma: it is located at ~ 2R⊙, hot but not extreme (log(T) ~ 7), producing unshifted lines, and displaying a very repetitive variability. A direct comparison with simulations yields an overall good agreement, with only a few further improvements needed.

The Of?p category was introduced more than 40 years ago to gather several Galactic stars with some odd properties. Since 2000, spectropolarimetry, high-resolution spectroscopy, long-term photometry, and X-ray observations have revealed their nature: magnetic oblique rotators - they all have magnetic fields that confine their winds. Several Of?p stars have now been detected in the Magellanic Clouds, likely the prototypes of magnetic massive stars at low metallicity. This contribution will present the most recent photometric, spectroscopic, and spectropolarimetric data, along with the first modeling of these objects.

Recent far-infrared (FIR) observations have revealed the presence of freshly formed dust with the masses exceeding 0.1 M⊙ in young remnants of core-collapse supernovae (CCSNe) such as SN 1987A and Cassiopeia A. Meanwhile, dust masses derived from near- to mid-infrared (N/MIR) observations of CCSNe a few years after explosions are on the order of 10−5–10−3M⊙. Here, we demonstrate that such small dust masses as seen from N/MIR observations would not necessarily reflect the formation history of dust but could be just limited by the luminosity of the SN that can heat up dust formed in the ejecta.

About one half of high-mass X-ray binaries host a Be star [an OB star with a viscous decretion (slowly outflowing) disk]. These Be/X-ray binaries exhibit two types of X-ray outbursts (Stella et al. 1986), normal X-ray outbursts (LX~1036−37 erg s−1) and occasional giant X-ray outbursts (LX > 1037 erg s−1). The origin of giant X-ray outbursts is unknown. On the other hand, a half of gamma-ray binaries have a Be star as the optical counterpart. One of these systems [LS I +61 303 (Porb = 26.5 d)] shows the superorbital (1,667 d) modulation in radio through X-ray bands. No consensus has been obtained for its origin. In this paper, we study a possibility that both phenomena are caused by a long-term, cyclic evolution of a highly misaligned Be disk under the influence of a compact object, by performing 3D hydrodynamic simulations. We find that the Be disk cyclically evolves in mildly eccentric, short-period systems. Each cycle consists of the following stages:

1. 1) As the Be disk grows with time, the initially circular disk becomes eccentric by the Kozai-Lidov mechanism.

2. 2) At some point, the disk is tidally torn off near the base and starts precession.

3. 3) Due to precession, a gap opens between the disk base and mass ejection region, which allows the formation of a new disk in the stellar equatorial plane (see Figure 1).

4. 4) The newly formed disk finally replaces the precessing old disk. Such a cyclic disk evolution has interesting implications for the long-term behavior of high energy emission in Be/X-ray and gamma-ray binaries.

Magnetic fields play an important role in shaping the circumstellar environment of hot, massive stars. Observational diagnostics give clues to the presence of magnetism across the entire electromagnetic spectrum. Infrared features can show more complex structure, indicating they may probe deeper opacities than optical features. Optical and infrared features mimic each other, with identical blue and red peak variations and identical peak velocity of material. These comparisons indicate the location of the infrared and optical emitting material is similar. Longer wavelength diagnostics are currently being developed and tested. IR spectroscopy is a viable tool to detect magnetic candidates in the Galactic center and star forming regions.

Turbulent transport and mixing generated by hydrodynamic instabilities triggered by rotation gradients are key mechanisms in the evolution of massive stars. We present here a summary of the progresses on shear-induced mixing obtained with numerical simulations, along with a new prescription for horizontal turbulence.

FASTWIND is a unified NLTE atmosphere/spectrum synthesis code originally designed (and frequently used) for the optical/IR spectroscopic analysis of massive stars with winds. Until the previous version (v10), the line transfer for background elements (mostly from the iron-group) was performed in an approximate way, by calculating the individual line-transitions in a single-line Sobolev or comoving frame approach, and by adding up the individual opacities and source functions to quasi-continuum quantities that are used to determine the radiation field for the complete spectrum (see Puls et al. 2005, A&A 435, 669, and updates).

We have now updated this approach (v11) and calculate, for all contributing lines (from elements H to Zn), the radiative transfer in the comoving frame, thus also accounting for line-overlap effects in an “exact” way. Related quantities such as temperature, radiative acceleration and formal integral have been improved in parallel. For a typical massive star atmospheric model, the computation times (from scratch, and for a modern desktop computer) are 1.5 h for the atmosphere/NLTE part, and 30 to 45 minutes (when not parallelized) for the formal integral (i.e., SED and normalized flux) in the ranges 900 to 2000 and 3800 to 7000 Å(Δλ = 0.03 Å).

We compare our new with analogous results from the alternative code CMFGEN (Hillier & Miller 1998, ApJ 496, 407, and updates), for a grid consisting of 5 O-dwarf and 5 O-supergiant models of different spectral subtype. In most cases, the agreement is very good or even excellent (i.e., for the radiative acceleration), though also certain differences can be spotted. A comparison with results from the previous, approximate method shows equally good agreement, though also here some differences become obvious. Besides the possibility to calculate the (total) radiative acceleration, the new FASTWIND version will allow us to investigate the UV-part of the spectrum in parallel with the optical/IR domain.

The formation of massive stars remains one of the most intriguing questions in astrophysics today. The main limitations result from the difficulty to obtain direct observational constraints on the formation process itself. In this context, the Carina High-contrast Imaging Project of massive Stars (CHIPS) aims to observe all 80+ O stars in the Carina nebula using the new VLT 2nd-generation extreme-AO instrument SPHERE. This instrument offers unprecedented imaging contrast allowing us to detect the faintest companions around massive stars. These novel observational constraints will help to discriminate between the different formation scenarios by comparing their predictions for companion statistics and properties.

We discuss the stellar and wind properties of 72 presumably single O-type giants, bright giants, and supergiants in the 30 Doradus region. This sample constitutes the largest and most homogeneous sample of such stars ever analyzed and offers the opportunity to test models describing their main-sequence evolution.

The role of massive stars in the Galactic evolution is crucial. During their lifetime these stars change the kinematics around them through stellar winds, affect the formation of new stars, ionise and chemically enrich the media with the final supernova explosion. But the census of both massive stars and their host clusters is still poor. We expect that still ~100 of galactic massive stellar clusters remains unknown (Hanson & Popescu, 2008).

We obtained VLT/X-shooter spectra of twelve candidate young massive stars previously selected by Hanson et al. (1997) in the giant Hii region M17. An analysis of their spectra using FASTWIND models (Puls et al. 2005) shows that they span a mass range of 6 - 20 M⊙. We identify the presence of gaseous and dusty disks around six sources based on emission lines in the spectrum and infrared continuum excess.