- Paper Title: Two ‘b’s in the Beehive: The Discovery of the First Hot Jupiters in an Open Cluster
- Authors: S. N. Quinn, R. J. White, D. W. Latham, et al.
- First Author’s Affiliation: Department of Physics & Astronomy, Georgia State University
Exoplanet studies over the last 15 years have found scores of gas giant planets that orbit close to their parent stars. Some of these ‘hot Jupiters’ are so close that they can disturb the stellar atmosphere itself. They have masses near to or exceeding that of Jupiter, but differ in that they orbit between 0.015 and 0.5 AU (by comparison, Jupiter orbits the Sun at approximately 5 AU). Studies have shown that at least 10% of FGK stars (stars similar in temperature to the sun) harbor hot Jupiters. Since most FGK stars form in open clusters, we might expect the number of hot Jupiters to be greater in these environments. This, however, does not seem to be the case: multiple radial velocity and transit searches have been carried out in clusters, and have failed to detect any planets in such environments.
Why is there a discrepancy between the overall frequency of planets and the frequency in clusters? One explanation might be the limited sample size of surveys. Alternatively, there may be a physical explanation, based on the fact that most hot Jupiters are found so close to their host star. This proximity implies that they have likely migrated inward after formation, since the leading theories of planet formation suggest that gas giants form beyond the snow line where water can freeze into ice. But in the environment of a cluster, solar-type stars do not possess massive enough disks to form gas giant planets at all, let alone support inward migration.
The authors of this paper, however, made a recent startling discovery that might alter our views on the characteristics and frequency of hot Jupiters, as well as the environments in which they form. Quinn et al. discovered two giant planets in the Beehive Cluster, the first known hot Jupiters in an open cluster.
Sample Selection and Observations
The Beehive Cluster was chosen because it is relatively nearby (170 pc), has approximately 1000 known members, a well-determined age, and an elevated metallicity. This last point is interesting because previous observations indicate that giant planet frequency is strongly correlated with host star metallicity. According to this correlation, the Beehive Cluster’s metallicity implies a giant planet frequency of 1 in 20 stars.
The authors utilized the radial velocity (RV) method for finding planets. Large and periodically varying RMS velocities indicate that these stars may in fact host planets. A massive exoplanet will cause a star to wobble due to its perturbing gravitational attraction. In other words the star and exoplanet are gravitationally attracted to one another, leading them to orbit around a point of mass central to both bodies. As such when the exoplanet is transiting on front of the host star or being occulted behind it, the star will have a velocity that is tangential to our line of sight (there will be no shift in wavelength). When the exoplanet is in quadrature (it makes a right angle with respect to the Sun) the star will either be moving toward us, and the starlight will be blue-shifted, or the star will be moving away from us, and the starlight will be red-shifted. Radial motion thus causes a Doppler shift in the spectrum of starlight, as the wavelengths increase or decrease depending on the motion of the star with respect to us, the observer, due to the position of the exoplanet.
The search excluded stars with known spectroscopic or visual companions, whose spectra would be dominated by the signal from their companion, masking out the weaker signal that would be induced by a planet. In addition, the authors preferentially targeted slowly rotating (to minimize the spectral line broadening effects rotation has on spectral lines), bright (to ensure spectral features strong enough to be reliably central) stars. After applying these selection criteria, 53 viable targets were included in the search. Given the giant planet frequency based on the cluster’s metallicity, 2-3 of these stars would be expected to host a giant gas planet.
The team conducted their survey using the 1.5 m reflector at the Fred L. Whipple Observatory on Mt. Hopkins in Arizona. In the end they were able to obtain 5-6 spectra of each of their 53 targets. Nightly observations of an RV standard star allowed them to improve their sensitivity to planets by correcting for three effects: (1) internal error (due to photon noise), (2) long-term instrumental error (night to night variations), and (3) astrophysical jitter (due to stellar activity).
The RMS velocities were calculated for the 53 stars in the sample. Figure 1 presents the radial velocities as a function of time for two stars in the cluster. The smooth, periodic oscillations are clear evidence that these two stars do in fact host large planets. These companions are clearly not binary stars in that they have masses equivalent to Jupiter. Based on the data presented in this paper, Pr0201b orbits orbits an F dwarf with a period of 4.4264 ± 0.0070 days and has a minimum mass of 0.540 ± 0.039 Mjup, while Pr0211b orbits a G dwarf with a period of 2.1451 ± 0.0012 days and has a minimum mass of 1.844 ± 0.064 Mjup. These uncertainties are much lower than other published results.
The discovery of two hot Jupiters in the Beehive Cluster confirms that short-period planets can exist in open clusters; this number is consistent with the expected number of planets in this metal-rich environment. If we assume that these planets did in fact migrate inward, then the time for this migration to occur must be less then the age of the cluster: 600 Myrs. This provides a very helpful constraint on migration times with another low uncertainty.
Uncertainties in planetary properties are often most limited by the large uncertainties in the properties of their host stars. While it is very difficult to measure accurate ages, radii, and masses of isolated, individual stars, these properties are much more easily and accurately inferred for stars in clusters. Consequently, the ages, radii, and masses of the planets around stars in clusters can be measured much more precisely. This remarkable discovery of two hot Jupiters in an open cluster is the first opportunity to study exoplanet characteristics with a much higher precision.
Nice post. One small correction: when a planet is transiting the host star and being occulted behind it, the star is moving tangentially to our line of sight. It is moving towards or away from us during times of quadrature.
Thank you Dr. Pepper for pointing that out! It has now been corrected for in the post.
Excellent post, Shannon. This was a really handy overview of a very exciting paper.
A few notes:
– The issue of completeness for open-cluster planet surveys is tricky. This paper by van Sanders & Gaudi presents the best study to date, demonstrating that not enough stars have been searched for transits to conclude that hot Jupiters are rare in clusters:
From the abstract: “Comparing these results to the frequency of short-period giant planets around field stars in both radial velocity and transit surveys, we conclude that there is no evidence to suggest that open clusters support a fundamentally different planet population than field stars, given the available data.”
– Hot Jupiters are typically defined as planets with Mp > 0.1 Mjup and P < 10 days. Given these limits, the occurrence rate is about 1%, not 10%. Check out Wright et al. (2012):
– The issue of when hot Jupiters migrate is an open question. If they have to migrate while there's a gas disk, then migration happens soon (< 10 Myr) after the star forms. However, if hot Jupiters are tossed in via gravitational interactions with other bodies, then this can, in principle, happen much later in the game. This is why it's so important to study hot Jupiters in young environments. The problem is that young stars are active (jittery) and rapidly rotating.
Keep up the great work writing about current science!