Spotting planets a long way from their stars is no easy proposition when you’re using radial velocity methods. The idea is to track the minute movement of the star as it is affected by an orbiting planet, which shows up as a Doppler shift in the data. What we’re actually seeing is the star and planet orbiting the center of gravity, an indirect method of detection that observes not the planet itself but the effects of the planet as it produces this variation in radial velocity.

The first exoplanets were detected this way, and the method has continued to produce new discoveries. But as a planet’s distance from its star increases, radial velocity becomes tricky to use. Now observation times become extended as the planet completes its longer orbit. We face the same issue with the transit method, which charts the drop in brightness as a planet moves across the face of its star as seen from Earth. Here, too, planets in distant orbits around their star are hard to detect because of the lengthy period of time between individual transits.

Because of these issues, we have little data on the occurrence rate of planets in wide orbits, and that is a problem for analyzing planet formation theories like core accretion, gravitational disk instability and planet migration. The long-term radial velocity data of a host star with a companion object beyond 10 AU shows an almost linear trend over a short observing period. Detection of such a trend is not in itself enough to identify the source of the RV signal.

rv_long

Image: Periodic change in the radial velocity of the primary star shown left is the strong signature of a planet. When the companion object is far way (farther than 10 AU), the orbital period is so long that the radial velocity shows a linear change, as shown in the right panel. (Credit: NAOJ).

Researchers from the Tokyo Institute of Technology have been trying to solve this problem by combining radial velocity methods with direct imaging. With the latter, the planet detection actually becomes easier as the distance between star and planet widens, because the light of the companion object is not swamped by proximity to the host star. The Tokyo team, using an instrument at Okayama Astrophysical Observatory (NAOJ) has targeted stars 1.5 to 5 times as massive as the Sun, looking not just for long period planets but also companion stars.

A long-term radial velocity trend flags such objects, but it’s the follow-up with direct imaging (in this case using the HiCIAO imager at the Subaru Telescope) that has paid off in an analysis of six intermediate-mass giant stars. All six of these stars showed a long-term radial velocity trend. The question thus becomes, is the source of the trend a companion star or a planet?

HiCIAO is a coronagraph that masks the light of the primary star to allow detection of fainter objects. Three of the observed stars — ? Hydra, HD 5608, and HS 109272 — show companion stars in this analysis, while around three others — ? Draconis, 18 Delphinus, and HD 14067 — companion objects more massive than 60 Jupiter masses can be excluded. The latter are considered candidates for hosting brown dwarfs, while 18 Delphinus is the most likely prospect for hosting a high-mass planet (~ 10-50 AU) that is below the current detection limit.

coronagraph

Image: Objects in the yellow circle were detected by the current study. White circles or round squares show the position of the primary star. The primary was masked during the observation to help the detection of the fainter objects nearby. (Credit: NAOJ)

Combining radial velocity with direct imaging has allowed the team to confirm that the companion objects they found were the cause of the long-term trend in the RV data. On a broader level, these methods may help us understand the distribution of planets around stars more massive than the Sun, as the paper on this work notes:

At Okayama Astrophysical Observatory (OAO), an RV survey targeting intermediate-mass giants (1.5–5 M?) has been conducted for over a decade (e.g., Sato et al. 2003). Sato et al. (2008) found that there is a difference between orbits of planets around intermediate-mass stars and around lower-mass FGK stars. Most planets around intermediate-mass stars have a semi-major axis larger than 0.6 AU, while FGK stars have shorter-period planets. Hence, it was suggested that the orbital distribution of exoplanets around intermediate-mass stars is different from that around solar-type stars.

Is the suggestion valid, or simply the result of our limited datasets? The paper continues:

In addition, the OAO survey detected long-term RV trends in several targets, which indicates the presence of distant companions around them… Identifying the companions that generate the RV trend can improve our knowledge of exoplanet populations for intermediate-mass stars, which are not well understood compared to solar-type stars.

This is not the only study that has combined direct imaging and radial velocity trends. In fact, a project called TRENDS (TaRgetting bENchmark objects with Doppler Spectroscopy), led by Justin Crepp (Notre Dame) has detected three low-mass stellar companions using these methods, along with a white dwarf and a brown dwarf companion, working with data on host stars ranging from F-class down to M-dwarfs. Clearly, direct imaging can be a benefit as we try to work out the source of RV trends that point to the existence of a distant companion.

The paper is Ryu et al., “High-contrast Imaging of Intermediate-mass Giants with Long-term Radial Velocity Trends,” Astrophysical Journal Vol. 825, No. 2 (12 July 2016). Abstract / preprint.

tzf_img_post