Open Access
Issue
A&A
Volume 619, November 2018
Article Number A81
Number of page(s) 16
Section Catalogs and data
DOI https://doi.org/10.1051/0004-6361/201833377
Published online 12 November 2018

© ESO 2018

Licence Creative Commons
Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1. Introduction

The spectrovelocimeter CORAVEL (COrrelation RAdial VELocities; Baranne et al. 1979) was installed on the Swiss 1-m telescope at the Observatory of Haute-Provence (OHP) from the late 1970s until its decommissioning in 2000. Amongst other programmes, it provided the radial-velocity (RV) measurements exploited in two statistical studies of binarity among the stars in the solar neighbourhood: the study of solar–type stars until G8, and its extension towards the K-type stars. A series of articles has been devoted to these programmes. The first (Duquennoy et al. 1991; Paper I hereafter) presented the radial-velocity measurements of the sample of F-G type stars; these data led to the orbital elements of several spectroscopic binaries (SBs), and to the statistical properties of solar-type binaries (Duquennoy & Mayor 1991; DM91 hereafter). Later, Halbwachs et al. (2003; Paper III hereafter) extended the statistical investigations to the K-type binaries with periods shorter than ten years, again on the basis of CORAVEL observations. This paper presented the parameters relevant for statistics, namely the periods, the semi-amplitudes, the mass ratios, and the orbital eccentricities of the spectroscopic binaries, excluding the other orbital elements. The long period K-type binaries were eventually studied by Eggenberger et al. (2004).

The present paper completes the series by presenting the radial velocity measurements and the full set of orbital elements that gave rise to Paper III. It will give the orbits we have discovered all the visibility they deserve, so that they are henceforth taken into account in statistical studies, such as that of Raghavan et al. (2010). Moreover, they will be available for the validation of the spectroscopic orbits derived from the Radial Velocity Spectrometer of the Gaia satellite Gaia Collaboration 2016). The CORAVEL programme is presented in Sect. 2, the RV catalogue is in Sect. 3, and the spectroscopic orbital elements are in Sect. 4. Section 5 is the conclusion.

Table 1.

Sample records of the RV measurement catalogues.

2. The CORAVEL survey of nearby K-type stars

The CORAVEL survey for nearby SBs was initiated in the early 1980s, although some stars (especially among the F–G types) had been measured before. The stars were taken from the second edition of the Catalogue of Nearby Stars (CNS; Gliese 1969) and from its supplement (Gliese & Jahreiss 1979). The stars discarded from the preliminary third version of the CNS (CNS3; Gliese & Jahreiss 1991) were kept in the observing runs. All stars were observed with CORAVEL from the Haute-Provence Observatory. Due to the location of the instrument and to its characteristics, only the stars as late as F7 and northern to −15° in declination were observed. Some stars with declination below −15° were observed, but they were not taken into account in Paper III. The programme was split according to the spectral types of the stars: the search for SBs amongst 288 F-G stars ended in December 1989, but the detection of SBs amongst the K-type stars was intensively performed until July 1993. After this date, the SBs were observed until 2000; at the same time, the RV of a few stars were still measured in order to confirm that it was constant.

3. Radial-velocity catalogues

3.1. CORAVEL individual measurements

The catalogue of the RV measurements provides 5413 measurements for 269 stars: 261 K-type stars and eight stars from the sample of DM91. These eight G-type stars were already in Paper I or another paper quoted in DM91, but they fulfil two conditions: they received enough additional RV measurements between 1991 and 2000 to significantly improve their spectroscopic orbit, and this new CORAVEL orbit was not published elsewhere. Moreover, the reduction of the CORAVEL observations was slightly improved, and the RV measurements are not exactly the same as in Paper I.

The format of the catalogue is presented in Table 1. Each record consists in the following data:

  • The number of the star in the CNS, followed by a letter designating the component, if any.

  • The epoch of the observation, given as a date with the year, the month, and the day, and also as barycentric Julian Day (BJD).

  • The RV, in km s−1.

  • The uncertainty of the RV.

  • The index of component (“1” for the primary, “2” for the secondary).

  • A flag “R” indicates the measurement was discarded from the calculation of the orbital elements.

The records are sorted by stars (from the smallest to the largest right ascension), and then by observation epochs.

3.2. Elodie individual measurements

The CORAVEL observations were not sufficiently accurate to derive valuable SB orbital elements for two stars of the sample, GJ 1069 and GJ 554. The latter of these two stars is even a constant velocity star when only the CORAVEL RVs are considered. Fortunately, RV measurements were performed thanks to the spectrograph Elodie of the 193 cm telescope at the Haute-Provence Observatory, and they are provided by the Elodie archive1. Fifteen Elodie RVs of GJ 1069 and 58 RVs of GJ 554 are presented in a separate file, with a slightly different format due to their accuracy. The uncertainties of the Elodie RVs were estimated as explained in Sect. 4.1. Sample records are presented at the end of Table 1.

3.3. The mean RV and the detection of SBs

The RV measurements were used to derive statistical information to decide whether a star is binary or not. These data are provided in Table A.1. The content is the following:

  • The identification of the star is the CNS number (GJ), as in the RV catalogue, and another identification, which is HD when it exists, otherwise BD, or HIP, or AG (Astronomische Gesellschaft catalogue). Three stars are designated only by the GJ identification; these stars are all visual secondary components, as indicated by letter “B” following their GJ number: GJ 57.1B, GJ 615.1B, and GJ 764.1B.

  • The BV colour index used to derive the CORAVEL RVs of the star. For the eight stars from the G-type sample, BV = 0.63 was assumed.

  • The mean RV, . When the star is a binary with known orbital elements, the RV of the barycentre is provided, as it is in Table A.2.

  • ϵ is the uncertainty of .

  • σRV is the standard deviation of the RV measurements.

  • E/I is the ratio of external to internal errors.

  • P(χ2) is the probability to obtain the χ2 of the RVs of the star, assuming that the RV is constant in reality.

  • N1 is the number of observations of the star.

  • Ntot is the number of RV measurements of the primary and of the secondary components.

  • ΔT is the time span of the observation.

  • The spectroscopic status is “CST” (ConStanT) when P(χ2) is more than 1%. Otherwise, it is SB1, or SB2 when the RV of the secondary component was measured. An “O” indicates that the orbital elements were derived, as explained in Sect. 4 hereafter. One star, GJ 554, has a constant CORAVEL RV, but an SB orbit derived by adding Elodie RV; this star is flagged “CSTO”. The status of two stars (GJ 893.2B and GJ 907.1) is “?” since only one RV measurement was obtained; these stars have both declinations around − 10°, and are too faint to be easily observable. However, they were only in the so-called “extended sample” of Paper III, and they were not relevant in the statistical investigations.

  • A flag indicates that the SB orbital elements of the star are in Table A.2, “*” when they are derived for the first time, and “+” otherwise. The velocity curves of the flagged stars are in Figs. 35.

The efficiency of the detection of the SBs depends on the time spans of the RV survey and on the numbers of RV measurements per star. The histograms of N1 and of Δ T are presented in Figs. 1 and 2, respectively. The stars with variable RV are counted apart from the others, since they received more observations when their variability was detected.

thumbnail Fig. 1.

Histogram of the number of observations. The shaded area refers to the stars that were not considered as variable and the white area to the spectroscopic binaries. The last bin represents all the stars with at least 21 observations; for the spectroscopic binaries, the count in this bin is 72.

Open with DEXTER

thumbnail Fig. 2.

Histogram of the time span of the observations. The shaded area refers to the stars that were not considered as variable and the white area to the spectroscopic binaries.

Open with DEXTER

It appears from Fig. 1 that a few stars received much less observations than the others. In addition to the two stars with only one observation already mentioned above, two stars have three observations, although one of them have a variable RV. This star is GJ 142, which was not taken into account in the binarity statistics because its declination is close to −20°. The other star is GJ 764.1B; it is difficult to observe since it is 5 arcsec away from its brighter companion GJ 764.1A, and it is only in the “extended” sample.

Half of the 209 constant RV stars received 11 observations or less. For the 269 stars in Table A.1, the median number of observations is 12.

The distribution of the time span, Fig. 2, also indicates that a few stars seem to have been less well observed than others. In addition to the two stars with one measurement and ΔT = 0, 3 stars were observed during less than 1500 days, although their RVs were variable. In fact, these stars are GJ 1124, GJ 343.1, and GJ 870, three short-perid binaries, and their observations were completed in a few years; they received enough RV measurements to derive their orbital elements, which are listed in Table A.2. The median time span is 3689 days for all the 270 stars, and 3410 days for the 209 constant RV stars.

thumbnail Fig. 3.

Spectroscopic orbits of the first 15 SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements and, for GJ 1069, the open triangles refer to the measurements obtained with Elodie; the Elodie RVs are shifted to the zero point of the CORAVEL measurements.

Open with DEXTER

thumbnail Fig. 4.

Spectroscopic orbits of the second set of 15 SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements and, for GJ 554, the open triangles refer to the measurements obtained with Elodie; the Elodie RVs are shifted to the zero point of the CORAVEL measurements.

Open with DEXTER

thumbnail Fig. 5.

Spectroscopic orbits of the 14 last SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements.

Open with DEXTER

4. Orbital elements of the spectroscopic binaries

4.1. Taking into account the Elodie RV measurements of GJ 1069 and GJ 554

The accuracy of the CORAVEL RVs precludes the derivation of relevant orbital elements for GJ 1069 and GJ 554, and it was necessary to take into account RVs provided by the Elodie archive. For that purpose, the uncertainty of the Elodie RVs must be estimated in order to assign them a reliable weight with respect to the CORAVEL RVs. For each star, the same weight was assigned to all the Elodie RV measurements and the SB orbital elements were derived from Elodie alone. The residuals of the RVs were calculated and the uncertainty of the RVs was then chosen so that the F2 estimator of the goodness-of-fit is zero. According to Stuart & Ord (1994) F2 is derived from theRV formula

(1)

where ν is the number of degrees of freedom and χ2 is the weighted sum of the squares of the differences between the predicted and the observed RVs, normalised with respect to their uncertainties. It was thus found that the uncertainty is 44 m s−1 for GJ 1069 and 14 m s−1 for GJ 554. The final elements were then derived taking into account simultaneously the RVs from Elodie and those of CORAVEL. A systematic offset between the two sets of measurements was derived with the SB solution. It is worth noticing that, although the CORAVEL RVs have weights much smaller than those of Elodie, they slightly improve the accuracy of the solution terms.

4.2. Presentation of the orbital elements

We used the RV measurements to derive the SB orbital elements for all stars sufficiently observed. The number of these stars is 45; they are flagged “O” in Table A.1. An SB orbit based on our CORAVEL RV measurements has already been published for three of them: GJ 1064B, GJ 692.1, and GJ 765.2; the orbits of the last two stars are even SB+VB orbits, since interferometric observations were also taken into account; the references of these three orbits are given in the footnotes of Table A.1. The SB orbital elements of the remaining 42 stars are listed in Table A.2. Since two stars are triple spectroscopic systems, this table contains 44 orbits.

The footnotes of Table A.2 indicate that 23 of these orbits were previously published on the basis of a part of our measurements or from RVs measured with a different instrument than CORAVEL. The 21 other orbits refer to new SBs.

The orbits of eight G-type stars mentioned in DM91 are included in Table A.2. DM91 provided a preliminary orbit for six of these stars, but the orbital elements are significantly improved there, thanks to additional RV measurements. The 44 SB orbits in Table A.2 are presented in Figs. 35.

5. Conclusion

We have drawn up a catalogue of 5413 RV measurements obtained with CORAVEL for 269 stars, 261 K-type dwarfs, and eight G-type dwarfs of the solar neighbourhood. These measurements were used to detect the SBs on which were based the statistical investigations of Paper III. We calculated the elements of 44 SB orbits, corresponding to 42 stellar systems. Twenty-one orbits, corresponding to 20 stellar systems, are the first orbits ever published for these stars.

All these data will be available through the VizieR service of the Centre de Donnée astronomique de Strasbourg (CDS). The SB orbits and the corresponding RV measurements will also be included in the on-line SB9 catalogue2

Acknowledgments

It is a pleasure to thank Salim Hillali for his contribution to the derivation of the orbits. The referee, Roger Griffin, indicated some relevant references. We have benefited during the entire period of these observations from the support of the Swiss National Foundation and Geneva University. We are particularly grateful to our technicians Bernard Tartarat, Emile Ischi, and Charles Maire for their dedication to that experiment for more than 20 years. We made use of Simbad, the database of the CDS and of the SB9 catalogue.


References

Appendix A: Tables of mean RVs and of orbital elements

Table A.1.

Average RV and the variability status.

Table A.2.

Orbital elements of the SBs.

All Tables

Table 1.

Sample records of the RV measurement catalogues.

Table A.1.

Average RV and the variability status.

Table A.2.

Orbital elements of the SBs.

All Figures

thumbnail Fig. 1.

Histogram of the number of observations. The shaded area refers to the stars that were not considered as variable and the white area to the spectroscopic binaries. The last bin represents all the stars with at least 21 observations; for the spectroscopic binaries, the count in this bin is 72.

Open with DEXTER
In the text
thumbnail Fig. 2.

Histogram of the time span of the observations. The shaded area refers to the stars that were not considered as variable and the white area to the spectroscopic binaries.

Open with DEXTER
In the text
thumbnail Fig. 3.

Spectroscopic orbits of the first 15 SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements and, for GJ 1069, the open triangles refer to the measurements obtained with Elodie; the Elodie RVs are shifted to the zero point of the CORAVEL measurements.

Open with DEXTER
In the text
thumbnail Fig. 4.

Spectroscopic orbits of the second set of 15 SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements and, for GJ 554, the open triangles refer to the measurements obtained with Elodie; the Elodie RVs are shifted to the zero point of the CORAVEL measurements.

Open with DEXTER
In the text
thumbnail Fig. 5.

Spectroscopic orbits of the 14 last SBs in Table A.2; the circles refer to the non-rejected CORAVEL RV measurements.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.