Free Access
Issue
A&A
Volume 573, January 2015
Article Number A3
Number of page(s) 6
Section Planets and planetary systems
DOI https://doi.org/10.1051/0004-6361/201424771
Published online 09 December 2014

© ESO, 2014

1. Introduction

Among the currently known ensemble of exoplanets, only a small fraction have been detected around giant stars, mainly because most of the radial velocity (RV) surveys have targeted solar-type stars thus far. However, such planets have properties that are very different from those orbiting main-sequence stars. In particular, giant planets are more common (~5% for MS host stars versus ~10% for giant host stars; Döllinger et al. 2009), there are few of them in close-in orbits (a 0.6 AU; Johnson et al. 2007), and they are significantly more massive than planets orbiting less evolved stars (Jones et al. 2014). Additionally, the fraction of brown dwarfs (BD) in relatively close-in orbits around giant stars also seems to be higher compared to less evolved host stars (Mitchell et al. 2013). These peculiar properties of exoplanets around giant stars give us information regarding the role of stellar mass on the formation of substellar objects and the effect of the stellar evolution in the orbital and physical properties of these systems (see discussion in Jones et al. 2014).

In 2009, we began the EXPRESS project (EXoPlanets aRound Evolved StarS; Jones et al. 2011), aimed at detecting planets around evolved (giant) stars via precision radial velocities, and to study their orbital properties. So far, we have discovered two planetary systems (Jones et al. 2013; 2014) plus the three substellar objects presented in this paper. Interestingly, three of these are super-Jupiters and two are in relatively close-in orbits (a 0.6 AU).

In this paper we present a planetary system orbiting the star HIP 67851, along with the discovery of a highly eccentric brown dwarf orbiting at ~2.5 AU from the star HIP 97233. The inner planet orbiting HIP 67851 is located at an orbital distance of 0.46 AU whereas the outer object has a much larger separation (a ≳ 4 AU). In Sect. 2, we describe the observations and the data reduction. In Sect. 3, we briefly present the radial velocity computation method. In Sect. 4, we summarize the main properties of the host stars. In Sect. 5, we present the orbital parameters of HIP 67851 b, c and HIP 97233 b. Finally, in Sect. 6, we present the summary and discussion.

2. Observations and data reduction

During the past five years we have collected ~15 spectra for each of the 166 targets in the EXPRESS program (Jones et al. 2011; 2013), using three different high-resolution optical spectrographs installed in Chile. Among these targets, we have followed up all of them showing RV variations above the ~30 m s-1 level, to determine whether they exhibit periodic variations that might be induced by orbiting planets. In particular, we took a total of 60 spectra for HIP 67851 and 41 spectra for HIP 97233, using the Fiber-fed Extended Range Optical Spectrograph (FEROS; Kaufer et al. 1999) and CHIRON (Tokovinin et al. 2013). Since these two stars are relatively bright, we obtained a signal-to-noise 100 with relatively short exposure times (~5 and 15 min for FEROS and CHIRON, respectively). The FEROS spectra were extracted using the FEROS data reduction system, while for CHIRON we used the reduction pipeline that is provided for the users. These two pipelines perform a standard reduction of echelle spectra (bias subtraction, flat-fielding correction, order tracing, extraction and wavelength calibration). For more details, see Jones et al. (2013; 2014).

3. Radial velocities

The method used to extract the radial velocities has been explained in detail in previous papers (Jones et al. 2013; 2014). Since we have used different instruments, we had to apply two different methods, namely, the simultaneous calibration method (Baranne et al. 1996) for FEROS data and the I2 cell technique (Butler et al. 1996) for CHIRON spectra. In the former case, we computed the cross-correlation function (CCF) between the observed spectra and a template, which corresponds to a high signal-to-noise ratio observation of the same star. This procedure is repeated independently for ~50 Å  chunks. We corrected the resulting velocity by the nightly drift, which is computed by cross-correlating the simultaneous calibration spectrum that is recorded in the sky fiber. Finally, we add the barycentric correction, which leads to the total final radial velocity at each observational epoch. The typical precision that we achieved for our targets is ~57 m s-1 (although it is possible to reach ~3 m s-1 in bright objects. See Jones et al. 2013). For the I2 cell observations, we modeled the observed spectrum of the star, which also contains the I2 absorption spectrum, in a similar fashion as presented in Butler et al. (1996). However, we use a simpler model for the instrumental point-spread-function, including only three Gaussians (a central profile of fixed width and varying height and two flanking profiles with varying height and width) and we applied this method to 5 Å  chunks. Although our model was optimized for the old Fiber Echelle spectrograph (which was replaced by CHIRON), it leads to a precision of ~5 m s-1 for CHIRON data, which is well below the RV noise introduced by stellar pulsations in giant stars (e.g., Sato et al. 2005).

Table 1

Stellar parameters of the host stars.

thumbnail Fig. 1

Position of HIP 67851 (open square) and HIP 97233 (filled circle) in the HR diagram. Different evolutionary tracks (Salasnich et al. 2000) for 1.7 M and 2.0 M are overplotted. The solid and dotted lines correspond to [Fe/H] = 0.0 and [Fe/H] = 0.32, respectively.

Open with DEXTER

thumbnail Fig. 2

Upper panel: radial velocity curve of HIP 67851. The red triangles and black dots correspond to CHIRON and FEROS velocities, respectively. One possible Keplerian solution with two giant planets is overplotted (black solid line). Lower panel: residuals from the Keplerian fit. The RMS around the fit is 9.4 m s-1.

Open with DEXTER

4. Stellar properties

The main stellar properties of HIP 67851 and HIP 97233 are listed in Table 1. The colors and parallaxes, as well as their spectral type, were taken from the Hipparcos catalog (Van Leeuwen 2007). The atmospheric parameters of these two stars were computed using the MOOG code (Sneden 1973), by matching the curve of growth of ~150 Fe i and ~20 Fe ii weak lines, as explained in Jones et al. (2011). Figure 1 shows the position of HIP 67851 (open square) and HIP 97233 (filled circle) in the HR diagram. For comparison, two evolutionary tracks for 1.7 M and 2.0 M from Salasnich et al. (2000) models are overplotted, which are the closest to the position of these stars. The solid and dotted lines correspond to [Fe/H] = 0.0 and [Fe/H] = 0.32, respectively. As can be seen, the two stars are ascending the red giant branch (RGB), although HIP 97233 is closer to the transition region between the end of the subgiant phase and the base of the RGB. Following the procedure described in Jones et al. (2011), we derived a mass of 1.63 ± 0.22 M and stellar radius of 5.92 ± 0.44 for HIP 67851. Similarly, we obtained a mass of 1.93 ± 0.11 M and radius of 5.34 ± 0.55 R for HIP 97233. For comparison, we obtained masses for these two stars using the method presented in da Silva et al. (2006)1. We used the spectroscopic values given in Table 1 and we corrected the V magnitudes with the AV values obtained from Arenou et al. (1992) extinction maps. In addition, we used the Girardi et al. (2002) models. We obtained a mass of 1.41 ± 0.18 M and radius of 5.95 ± 0.36 R for HIP 67851, and a mass of 1.84 ± 0.14 M and radius of 5.20 ± 0.50 R for HIP 97233. In both cases, the mass and radius derived by the two methods agree within one sigma.

thumbnail Fig. 3

Bisector velocity span and FWHM of the CCF versus FEROS radial velocities of HIP 67851 (upper and lower panel, respectively).

Open with DEXTER

5. Orbital parameters

5.1. HIP 67851  b, c

Figure 2 shows the RV measurements of HIP 67851. The black dots and red triangles correspond to FEROS and CHIRON velocities, respectively. The RV variations are listed in Tables A.1 and A.2. The solid line represents one possible Keplerian solution with an outer planet with MP = 6.3MJ, P = 2200 days and zero eccentricity. The orbital parameters of the inner planet (HIP 67851 b) are listed in Table 2. The orbital fit was computed using the systemic console (Meschiari et al. 2009). The uncertainties were computed using a bootstrap randomization method. The RMS around the fit is 9.4 m s-1. As can be seen, although the orbital parameters of the inner planets are well determined, further follow up is needed to fully constrain the orbital period of the outer object, which probably has a mass in the planetary regime. The best Keplerian solution for HIP 67851 b yields an orbital distance of 0.46 AU, meaning that it is one of the few known planets orbiting interior to 0.5 AU around giant stars2. We tested the nature of the periodic RV variations observed in HIP 67851, using the Hipparcos photometry of this star. This dataset is comprised of 75 high-quality measurements. No significant periodic signal is observed in the data and the RMS is only 0.009 mag. In addition, we performed a line profile variation analysis. Figure 3 shows the bisector velocity span (BVS) and the full width at half maximum (FWHM) of the CCF versus the measured radial velocities (upper and lower panel, respectively). These values were computed according to the procedure explained in Jones et al. (2013). Clearly there is no correlation between these two quantities and the RV measurements. Finally, Fig. 4 shows the chromospheric activity index (S-index) as a function of the HIP 67851 RVs. These values were computed according to the procedure described in Jenkins et al. (2008; 2011). No obvious correlation is observed between these quantities. Based on the results of these three tests, along with a failed periodogram search for any frequencies present in those data, we conclude that the RV variations observed in HIP 67851 are most likely induced by two substellar objects.

thumbnail Fig. 4

Chromospheric activity indicator (S-index) versus the HIP 67851 FEROS radial velocities.

Open with DEXTER

Table 2

Orbital parameters of HIP 67851 b and HIP 97233 b.

thumbnail Fig. 5

Upper panel: radial velocity curve for HIP 97233. The red triangles and black dots correspond to CHIRON and FEROS velocities, respectively. The best Keplerian solution is overplotted (black solid line). Lower panel: residuals from the Keplerian fit. The RMS around the fit is 10.1 m s-1.

Open with DEXTER

5.2. HIP 97233 b

Figure 5 shows the RV measurements of HIP 97233. The black dots and red triangles correspond to FEROS and CHIRON velocities, respectively. The solid line represents the best Keplerian fit. The orbital parameters are listed in Table 2. The RMS around the fit is 10.1 m s-1, which is consistent with the RV noise induced by stellar oscillations in giant stars (e.g., Hekker et al. 2006). The RV variations are listed in Tables A.3 and A.4.

As for HIP 67851, we investigated different possible mechanisms that might be causing the RV variations observed in HIP 97233. We analyzed the Hipparcos photometric data of this star, which consists of a total of 46 high quality observations. There is no significant periodic signal in the data and the RMS is below 0.01 mag. Such a low variability is not expected to mimic a signal as large as the signal observed in the HIP 97233 data (Hatzes 2002). Additionally, we performed a line profile analysis. The results from the BVS and FWHM variations are plotted in Fig. 6 (upper and lower panel, respectively). No obvious correlation is observed between these values and RV variations. Finally, Fig. 7 shows the S-index versus the observed radial velocities of HIP 97233. Clearly, there is no correlation with the radial velocities. Based on these results we conclude that the most likely explanation for the periodic RV variations observed in HIP 97233 is due to the presence of a substellar object.

thumbnail Fig. 6

Bisector velocity span and FWHM of the CCF versus FEROS radial velocities of HIP 97233 (upper and lower panel, respectively).

Open with DEXTER

thumbnail Fig. 7

Chromospheric activity indicator (S-index) versus the HIP 97233 FEROS radial velocities.

Open with DEXTER

6. Summary and discussion

In this paper we present the discovery of two substellar systems found around the giant stars HIP 67851 and HIP 97233. The radial velocity variations of HIP 67851 are most likely explained by the presence of two planetary mass objects. The inner planet in the system has the following orbital parameters: MP sini = 1.44MJ, P = 88.8 days and e = 0.09. HIP 67851 b is one of the few planets orbiting interior to ~0.5 AU around giant stars. The orbital properties of the outer object are still uncertain, since only a fraction of its orbital period has been covered. However, it is probably a super-Jupiter or a BD in a wide orbit (a ≳ 4 AU). This new discovery confirms the finding by Schlaufman & Winn (2013) and Jones et al. (2014) since HIP 67851 b has a nearly circular orbit (e ≲ 0.2) and a semi-major axis 0.9 AU. Moreover, if HIP 67851 c is actually a super-Jupiter, it would also confirm that most of the planets around giant stars are super-planets.

On the other hand, the RV curve of HIP 97233 shows a periodic signal that is most likely induced by a BD with the following orbital parameters: MP sini = 20.0MJ, P = 1058.8 days and e = 0.61. With this new discovery, there are ~10 known BDs around giant stars, with HIP 97233 b being the most eccentric of these. Although this number is small, it is similar to the number of known BDs orbiting interior to ~4 AU around solar-type stars. However, since the number of giant stars targeted by RV surveys is relatively small, it seems evident that the fraction of BDs around giant stars is significantly higher than around solar-type stars (Mitchell et al. 2013). The reason for this difference in the detection rate is not because of an observational bias, rather it is probably explained by either the role of the host star mass (since the giant host stars are mainly intermediate-mass stars) or by the effect of the stellar evolution. In the former case, the mass and content of the protoplanetary disk might strongly affect the efficiency of planet formation and the subsequent gas accretion. Thus, it is possible that intermediate-mass stars have more massive and dense disks, from which massive planets and BD are more efficiently formed (Lovis & Mayor 2007). On the other hand, the mass of these super-planets might be explained by the evolution of the host star. In this scenario, a giant planet might be able to accrete a significant amount of mass from an enhanced stellar wind during the giant branch phase, eventually exceeding the deuterium burning limit and becoming a BD (Jones et al. 2014). It is also notable that the metallicity of HIP 97233 is very high ([Fe/H] = 0.29 ± 0.13 dex). It might thus be possible that its brown dwarf companion might have formed through core accretion (see Jenkins et al. 2009), in a process similar to what is thought to be the dominant formation mechanism for planets. If this is the case, then we would expect to detect more massive objects orbiting the most metal-rich stars since these disks form higher mass cores more efficiently, giving rise to a deficit of low-mass objects orbiting the most metal-rich stars (Jenkins et al. 2013). Finally, because of the interaction with either the disk before the gas is dissipated or with the stellar envelope during the post-MS phase, these BDs might have migrated from beyond ~4 AU to their current position. In this scenario, the primordial fraction of BDs could be the same as for solar-type host stars.


2

Along with the transiting planets Kepler-56  b, c (Huber et al. 2013), Kepler-91 b (Lillo-Box et al. 2014a,b; Barclay et al. 2014) and KOI-1299  b (Ciceri et al. 2014; Ortiz et al. 2014). In addition, HIP 13044 b was shown to likely not be a real planet (Jones & Jenkins 2014).

Acknowledgments

M.J. acknowledges financial support from Fondecyt project #3140607. J.J and P.R. acknowledge funding from CATA (PB06, Conicyt). P.R. acknowledges support from Fondecyt project #1120299 and Anillo ACT1120. We acknowledge the anonymous referee for useful comments. This research has made use of the SIMBAD database and the VizieR catalogue access tool, operated at CDS, Strasbourg, France.

References

Appendix A: Radial velocity tables

Table A.1

FEROS radial velocity variations of HIP 67851.

Table A.2

CHIRON radial velocity variations of HIP 67851.

Table A.3

FEROS radial velocity variations of HIP 97233.

Table A.4

CHIRON radial velocity variations of HIP 97233.

All Tables

Table 1

Stellar parameters of the host stars.

Table 2

Orbital parameters of HIP 67851 b and HIP 97233 b.

Table A.1

FEROS radial velocity variations of HIP 67851.

Table A.2

CHIRON radial velocity variations of HIP 67851.

Table A.3

FEROS radial velocity variations of HIP 97233.

Table A.4

CHIRON radial velocity variations of HIP 97233.

All Figures

thumbnail Fig. 1

Position of HIP 67851 (open square) and HIP 97233 (filled circle) in the HR diagram. Different evolutionary tracks (Salasnich et al. 2000) for 1.7 M and 2.0 M are overplotted. The solid and dotted lines correspond to [Fe/H] = 0.0 and [Fe/H] = 0.32, respectively.

Open with DEXTER
In the text
thumbnail Fig. 2

Upper panel: radial velocity curve of HIP 67851. The red triangles and black dots correspond to CHIRON and FEROS velocities, respectively. One possible Keplerian solution with two giant planets is overplotted (black solid line). Lower panel: residuals from the Keplerian fit. The RMS around the fit is 9.4 m s-1.

Open with DEXTER
In the text
thumbnail Fig. 3

Bisector velocity span and FWHM of the CCF versus FEROS radial velocities of HIP 67851 (upper and lower panel, respectively).

Open with DEXTER
In the text
thumbnail Fig. 4

Chromospheric activity indicator (S-index) versus the HIP 67851 FEROS radial velocities.

Open with DEXTER
In the text
thumbnail Fig. 5

Upper panel: radial velocity curve for HIP 97233. The red triangles and black dots correspond to CHIRON and FEROS velocities, respectively. The best Keplerian solution is overplotted (black solid line). Lower panel: residuals from the Keplerian fit. The RMS around the fit is 10.1 m s-1.

Open with DEXTER
In the text
thumbnail Fig. 6

Bisector velocity span and FWHM of the CCF versus FEROS radial velocities of HIP 97233 (upper and lower panel, respectively).

Open with DEXTER
In the text
thumbnail Fig. 7

Chromospheric activity indicator (S-index) versus the HIP 97233 FEROS radial velocities.

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.