Detecting Transits of Planetary Companions to Giant Stars

Of the approximately 350 extrasolar planets currently known, of order 10% orbit evolved stars with radii R >~ 2.5 R_sun. These planets are of particular interest because they tend to orbit more massive hosts, and have been subjected to variable stellar insolation over their recent histories as their primaries evolved off the main sequence. Unfortunately, we have limited information about the physical properties of these planets, as they were all detected by the radial velocity method and none have been observed to transit. Here we evaluate the prospects for detecting transits of planetary companions to giant stars. We show that several of the known systems have a priori transit probabilities of >~ 10%, and about one transiting system is expected for the sample of host stars with R >= 2.5 R_sun. Although the transits are expected to have very small amplitudes (few x 10^-4) and long durations (> 50 hrs), we argue that the difficulty with detecting these signals in broadband light is one of systematic errors and practicality rather than photon noise, even for modest aperture ~1m telescopes. We propose a novel method that may overcome these difficulties, which uses narrow-band measurements to isolate the thin ring of chromospheric emission expected at the limb of giant stars. The transit signals in these narrow bands are expected to be larger in magnitude and briefer in duration than in broad-band emission, and thus alleviating many of the difficulties with transit detection in broad-band emission. Finally, we point out that it may be possible to discover planetary companions to giant stars using Kepler, provided that a sufficient number of such targets are monitored.

💡 Research Summary

The paper addresses a conspicuous gap in exoplanet science: none of the roughly 350 known planets orbiting evolved stars (R ≳ 2.5 R☉) have been observed in transit. These planets, discovered exclusively by radial‑velocity (RV) surveys, are of high interest because their host stars are typically more massive than solar‑type stars and have undergone significant luminosity evolution, exposing the planets to changing irradiation histories. The authors first quantify the a‑priori transit probability for each known giant‑star system using the simple geometric expression Pₜᵣₐₙₛᵢₜ ≈ (R★ + Rₚ)/a. While the large stellar radii increase the probability, the depth of a transit scales as (Rₚ/R★)² and therefore drops to a few × 10⁻⁴ for a Jupiter‑sized planet transiting a 10 R☉ star. Moreover, the orbital periods are long (a ≈ 1–3 AU), giving orbital velocities of only a few km s⁻¹ and resulting in transit durations of 50–100 hours.

Statistically, the authors find that several individual systems have Pₜᵣₐₙₛᵢₜ > 10 %, and that, for the whole sample of stars with R ≥ 2.5 R☉, the expected number of transiting planets is about one. The principal obstacle to detecting such shallow, long‑duration events in broadband photometry is not photon noise but systematic errors: atmospheric transparency variations, flat‑field imperfections, and red (time‑correlated) noise that can easily mask a 10⁻⁴ signal even with a modest 1‑m telescope.

To overcome these limitations, the authors propose a novel narrow‑band technique that exploits the chromospheric emission ring that forms at the limb of giant stars. In lines such as Ca II H&K or Hα, the chromosphere emits strongly in a thin annulus whose radial thickness ΔR is only ~1 % of the stellar radius. When a planet occults this bright ring, the relative flux decrement is amplified by roughly (Rₚ/ΔR)², potentially increasing the observable signal by an order of magnitude or more compared with broadband light. Because the annulus is thin, the effective transit of the chromospheric ring lasts only a few hours, dramatically reducing the exposure to long‑term systematics. Simulations indicate that a 1–2 m telescope equipped with a high‑resolution spectrograph and a ~1 nm narrow filter could achieve a signal‑to‑noise ratio sufficient for a ≥5σ detection of the amplified event.

The paper also discusses the prospects for space‑based missions. The Kepler field contains several thousand giant stars; if a sufficient number are monitored continuously, the modest a‑priori probabilities translate into a realistic chance of catching at least one transit. However, Kepler’s relatively large pixel scale and background blending, together with intrinsic variability of giant stars, necessitate careful target selection and detrending.

In summary, while transits of planets around giant stars are intrinsically shallow and long, they are not beyond reach. The authors demonstrate that (1) the geometric probability is non‑negligible, (2) broadband detection is limited mainly by systematic noise, and (3) a targeted narrow‑band approach that isolates chromospheric limb emission can boost the signal and shorten the effective event, making ground‑based detection feasible. The methodology opens a pathway to directly measure radii and densities of planets that have survived stellar evolution, and it suggests that upcoming or archival space‑based photometric surveys could also uncover such rare but scientifically valuable systems.