Abstract. New radial velocities of α Cen A & B obtained in the framework the Anglo-Australian Planet Search programme as well as in the CORALIE programme are added to those by Endl et al. (2001) to improve the precision of the orbital parameters. The resulting masses are 1.105 ± 0.0070 M and 0.934 ± 0.0061 M for A and B respectively. The factors limiting how accurately these masses can be derived from a combined visual-spectroscopic solution are investigated. The total effect of the convective blueshift and the gravitational redshift is also investigated and estimated to differ by 215 ± 8 m s −1 between the components. This suggests that the difference in convective blueshift between the components is much smaller than predicted from current hydrodynamical model atmosphere calculations.

New astrometric–spectroscopic orbital solutions for the single‐line K‐giant binaries β Reticuli (P≈ 5.2 yr, e= 0.3346 ± 0.0004) and ν Octantis (P≈ 2.9 yr, e= 0.2358 ± 0.0003) have been derived based on high‐precision spectroscopic radial velocities (RVs) and the Hipparcos astrometry. For the case of ν Oct, the simultaneous solution is particularly robust and an inclination of i= 70.8 ± 0.9° has been derived. This is one of the most precise inclinations yet calculated based on a spectroscopic solution and the Hipparcos astrometry. We have also discovered low‐amplitude periodic behaviour in the residuals of the orbital solution for ν Oct. This RV perturbation has a semi‐amplitude of 50 m s−1 and a 418‐d period which is coherent over several years. The RV curve of the perturbation is apparently in resonance with that of the binary: every second maximum of the binary coincides with every fifth minimum of the perturbation, hence the periods have the simple ratio 5:2. The possible causes of such a perturbation are rotational modulation of surface phenomenon, pulsations or an orbiting body. We have assessed these alternatives in terms of the suspected photometric stability (Hp= 3.8981 ± 0.0004), a lack of evidence of other RV periodicities, no correlation of cross‐correlation function bisectors with the residual velocities, no compelling evidence of wavelength dependency for the amplitude or relative phase of the perturbation, our bounds on the rotational period of the primary star and the need for long‐lived relatively fixed surface features. The results of these analyses lack consistency with both rotational modulation and pulsations and so imply that a planetary mass is a realistic cause. The planet hypothesis, however, is strongly constrained and challenged by our precise binary orbit. The hypothetical planet would have an orbit (e≈ 0.1, a3≈ 1.2 au) about mid‐way between the stars whose periastron distance is only 1.9 au. This orbit, supposedly in resonance with the binary system, appears to be highly unlikely based on current planet formation and orbit‐stability expectations. Without knowing the cause of the perturbation, we cannot be certain if the suspected RV and hence period resonance are merely coincidental or not. Establishing the true cause of the perturbation requires renewed observation of the system, re‐assessment of the possible resonance if this is redetected and the acquisition of similar and additional precise diagnostic parameters with respect to each of the possible causative mechanisms.

No abstract

Accurate orbital parameters have been obtained for the bright southern single-lined spectroscopic binary star ζ TrA, with the 1-m telescope and Hercules spectrograph at Mt John University Observatory. More than 200 observations have been collected over five months of observation between 2003 March and August. They cover about 12 complete periods (P ≈ 13 d). High-precision relative radial velocities have been measured by the cross-correlation technique with HRSP, a dedicated reduction software package developed in the Department of Physics and Astronomy, University of Canterbury. An iterative linear least-squares method has been used to improve the initial set of orbital parameters estimated from the geometrical properties of the radial-velocity curve. The final rms error of the least-squares fit is only about 14 m s −1 so the error bars on the orbital parameters are extremely small. As a particularly interesting result, a very low eccentricity of e = 0.014 42 ± 0.000 21 has been derived, which proves that the orbit is definitely elliptical and not circular as has been suggested in the literature. The orbital period has also been improved by including some historical observations so that the best value is P = 12.975 780 ± 0.000 046 d. The mass of the companion is in the range 0.09-0.45 M , which covers the spectral types between M1 V and M7 V, on the assumption that the companion is a main-sequence star. In that case the orbital inclination has to be larger than about 14 • .

The 'holy grail' in planet hunting is the detection of an Earth-analogue: a planet with similar mass as the Earth and an orbit inside the habitable zone. If we can find such an Earth-analogue around one of the stars in the immediate solar neighbourhood, we could potentially even study it in such great detail to address the question of its potential habitability. Several groups have focused their planet detection efforts on the nearest stars. Our team is currently performing an intensive observing campaign on the α Centauri system using the High Efficiency and Resolution Canterbury University Large Échelle Spectrograph (HERCULES) at the 1 m McLellan telescope at Mt John University Observatory in New Zealand. The goal of our project is to obtain such a large number of radial velocity (RV) measurements with sufficiently high temporal sampling to become sensitive to signals of Earth-mass planets in the habitable zones of the two stars in this binary system. Over the past few years, we have collected more than 45 000 spectra for both stars combined. These data are currently processed by an advanced version of our RV reduction pipeline, which eliminates the effect of spectral cross-contamination. Here we present simulations of the expected detection sensitivity to low-mass planets in the habitable zone by the HERCULES programme for various noise levels. We also discuss our expected sensitivity to the purported Earth-mass planet in a 3.24-day orbit announced by Dumusque et al. (2012).