As an exoplanet transits its host star, some of the light from the star is absorbed by the atoms and molecules in the planets atmosphere, causing the planet to seem bigger; plotting the planets observed size as a function of the wavelength of the light produces a transmission spectrum 1 . Measuring the tiny variations in the transmission spectrum, together with atmospheric modelling, then gives clues to the properties of the exoplanets atmosphere. Chemical species composed of light elementssuch as hydrogen, oxygen, carbon, sodium and potassiumhave in this way been detected in the atmospheres of several hot giant exoplanets [2][3][4][5] , but molecules composed of heavier elements have thus far proved elusive. Nonetheless, it has been predicted that metal oxides such as titanium oxide (TiO) and vanadium oxide occur in the observable regions of the very hottest exoplanetary atmospheres, causing thermal inversions on the dayside 6, 7 . Here we report the detection of TiO in the atmosphere of the hot-Jupiter planet WASP-19b. Our combined spectrum, with its wide spectral coverage, reveals the presence of TiO (to a confidence level of 7.7σ), a strongly scattering haze (7.4σ) and sodium (3.4σ), and confirms the presence of water (7.9σ) in the atmosphere 5,8 .Hot Jupiters are gas-giant exoplanets with sizes like that of Jupiter but much shorter orbital periods. WASP-19b is the shortest-period hot Jupiter to be discovered so far 9 , and has an excessively bloated radius, owing to the extreme radiation that it receives from its host star; as a result of this radiation, the planets effective temperature is more than 2,000 K (obtained via secondaryeclipse measurements 10 ). It is thought that high atmospheric temperatures imply the presence of metal oxides such as TiO, but despite extensive searches 11, 12 a definitive detection of metal oxides in exoplanetary atmospheres has proved elusive.We observed three transits of European Southern Observatorys Very Large Telescope (VLT), using the low-resolution FORS2 spectrograph. By using three of FORS2s grisms600B (blue), 600RI (green) and 600z (red), thereby covering the entire visible-wavelength domain (0.431.04 µm)together with the multi-object spectroscopy configuration, we were able to obtain relatively high-resolution, precise, broadband transmission spectra. Such results were made possible through optimized observing strategies 13 and careful design of the observing mask used for the multi-object observations: this has slits about 30 wide, which minimized differential losses owing to variations in telescope guiding and seeing conditions. The observations presented here were made between 11 November 2014 and 29 February 2016.For each set of observations, we obtained a series of spectra for the main target (WASP-19), as well as for several comparison stars. After standard data-reduction steps, we integrated those spectra for the largest common wavelength domain and 10-nm bins, to produce the 'white' and 'spectrophotometric' light curves, respectively. To correct for the imp...

Context. Barium and S stars without technetium are red giants suspected of being all members of binary systems. Aims. This paper provides both long-term and revised, more accurate orbits for barium and S stars adding to previously published ones. The sample of barium stars with strong anomalies comprise all such stars present in the Lü et al. catalogue. Methods. Orbital elements are derived from radial velocities collected from a long-term radial-velocity monitoring performed with the HERMES spectrograph mounted on the Mercator 1.2 m telescope. These new measurements were combined with older, CORAVEL measurements. With the aim of investigating possible correlations between orbital properties and abundances, we collected as well an as homogeneous as possible set of abundances for barium stars with orbital elements. Results. We find orbital motion for all barium and extrinsic S stars monitored. We obtain the longest period known so far for a spectroscopic binary involving an S star, namely 57 Peg with a period of the order of 100 -500 yr. We present the mass distribution for the barium stars, which ranges from 1 to 3 M , with a tail extending up to 5 M in the case of mild barium stars. This high-mass tail comprises mostly high-metallicity objects ([Fe/H] ≥ −0.1). Mass functions are compatible with WD companions whose masses range from 0.5 to 1 M . Strong barium stars have a tendency to be found in systems with shorter periods than mild barium stars, although this correlation is rather lose, metallicity and WD mass playing a role as well. Using the initial -final mass relationship established for field WDs, we derived the distribution of the mass ratio q = M AGB,ini /M Ba (where M AGB,ini is the WD progenitor initial mass, i.e., the mass of the system former primary component) which is a proxy for the initial mass ratio (the more so, the less mass the barium star has accreted). It appears that the distribution of q is highly non uniform, and significantly different for mild and strong barium stars, the latter being characterized by values mostly in excess of 1.4, whereas mild barium stars occupy the range 1 -1.4. Conclusions. The orbital properties presented in this paper pave the way for a comparison with binary-evolution models.

The orbital parameters of binaries at intermediate periods (10 2 -10 3 d) are difficult to measure with conventional methods and are very incomplete. We have undertaken a new survey, applying our pulsation timing method to Kepler light curves of 2224 main-sequence A/F stars and found 341 non-eclipsing binaries. We calculate the orbital parameters for 317 PB1 systems (single-pulsator binaries) and 24 PB2s (double-pulsators), tripling the number of intermediate-mass binaries with full orbital solutions. The method reaches down to small mass ratios q ≈ 0.02 and yields a highly homogeneous sample. We parametrize the mass-ratio distribution using both inversion and MCMC forward-modelling techniques, and find it to be skewed towards low-mass companions, peaking at q ≈ 0.2. While solar-type primaries exhibit a brown dwarf desert across short and intermediate periods, we find a small but statistically significant (2.6σ) population of extreme-mass-ratio companions (q < 0.1) to our intermediatemass primaries. Across periods of 100 -1500 d and at q > 0.1, we measure the binary fraction of current A/F primaries to be 15.4% ± 1.4%, though we find that a large fraction of the companions (21% ± 6%) are white dwarfs in post-mass-transfer systems with primaries that are now blue stragglers, some of which are the progenitors of Type Ia supernovae, barium stars, symbiotics, and related phenomena. Excluding these white dwarfs, we determine the binary fraction of original A/F primaries to be 13.9% ± 2.1% over the same parameter space. Combining our measurements with those in the literature, we find the binary fraction across these periods is a constant 5% for primaries M 1 < 0.8 M , but then increases linearly with log M 1 , demonstrating that natal discs around more massive protostars M 1 1 M become increasingly more prone to fragmentation. Finally, we find the eccentricity distribution of the main-sequence pairs to be much less eccentric than the thermal distribution.

Smoothed particle hydrodynamics (SPH) is used to estimate accretion rates of mass, linear and angular momentum in a binary system where one component undergoes mass loss through a wind. Physical parameters are chosen such as to model the alleged binary precursors of barium stars, whose chemical peculiarities are believed to result from the accretion of the wind from a companion formerly on the asymptotic giant branch (AGB). The binary system modelled consists of a 3M ⊙ AGB star (losing mass at a rate 10 −6 M ⊙ y −1 ) and a 1.5M ⊙ star on the main sequence, in a 3 AU circular orbit. Three-dimensional simulations are performed for gases with polytropic indices γ = 1, 1.1 and 1.5, to bracket more realistic situations that would include radiative cooling. Mass accretion rates are found to depend on resolution and we estimate typical values of 1-2% for the γ = 1.5 case and 8% for the other models. The highest resolution obtained (with 400k particles) corresponds to an accretor of linear size ≈ 16R ⊙ . Despite being (in the γ = 1.5 case) about ten times smaller than theoretical estimates based on the Bondi-Hoyle prescription, the SPH accretion rates remain large enough to explain the pollution of barium stars. Uncertainties in the current SPH rates remain however, due to the simplified treatment of the wind acceleration mechanism, as well as to the absence of any cooling prescription and to the limited numerical resolution.Angular momentum transfer leads to significant spin up of the accretor and can account for the rapid rotation of HD 165141, a barium star with a young white dwarf companion and a rotation rate unusually large among K giants.In the circular orbit modelled in this paper, hydrodynamic thrust and gravitational drag almost exactly compensate and so the net transfer of linear momentum is nearly zero. For small but finite eccentricities and the chosen set of parameters, the eccentricity tends to decrease.

Stars are generally spherical, yet their gaseous envelopes often appear non-spherical when ejected near the end of their lives. This quirk is most notable during the planetary nebula phase when these envelopes become ionized. Interactions among stars in a binary system are suspected to cause the asymmetry. In particular, a precessing accretion disk around a companion is believed to launch point-symmetric jets, as seen in the prototype Fleming 1. Our discovery of a post common-envelope binary nucleus in Fleming 1 confirms that this scenario is highly favorable. Similar binary interactions are therefore likely to explain these kinds of outflows in a large variety of systems.Planetary nebulae (PNe) are thought to represent the transitory phase of the end of the lives of solar-like stars. The mass-loss mechanisms at play during the late stages of stellar evolution that produce the observed shapes of planetary nebulae have been a matter of debate in the last two decades (1). The leading paradigm to produce the most extreme nebular morphologies is evolution in an interacting binary system (2-4), in particular common-envelope (CE) evolution -the dramatic outcome of unstable mass transfer resulting in a binary system with a greatly reduced orbital period (P<~1 day for PNe). Despite recent detections of multiple post common-envelope binary central stars (5-7), there are as yet no clear-cut examples of binaries actively shaping their surrounding planetary nebulae. A handful of post-CE nebulae are known to be oriented in agreement with the orbital inclination of the binaries that ejected them (8) -as would be expected. However, we do not yet have any inkling how a particular binary configuration gives rise to a specific fundamental nebula shape. An alternative approach to tackle this difficult