Giant Planets in Open Clusters S.N. Quinn 1 , 3 , R.J. White 1 , D.W. - - PDF document

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold

Giant Planets in Open Clusters

S.N. Quinn1,3, R.J. White1, D.W. Latham2 Talk given at OHP-2015 Colloquium

1Department of Physics & Astronomy, Georgia State University, 25 Park Place NE Suite 605, Atlanta, GA 30316,

USA (quinn@astro.gsu.edu)

2Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA 3National Science Foundation Graduate Research Fellow

Abstract Two decades after the discovery of 51 Peg b, more than 200 hot Jupiters have now been con- firmed, but the details of their inward migration remain uncertain. While it is widely accepted that short period giant planets could not have formed in situ, several different mechanisms (e.g., Type II migration, planet-planet scattering, Kozai-Lidov cycles) may contribute to shrinking planetary or- bits, and the relative importance of each is not well-constrained. Migration through the gas disk is expected to preserve circular, coplanar orbits and must occur quickly (within ∼10 Myr), whereas multi-body processes should initially excite eccentricities and inclinations and may take hundreds

• f millions of years. Subsequent evolution of the system (e.g., orbital circularization and inclina-

tion damping via tidal interaction with the host star) may obscure these differences, so observing hot Jupiters soon after migration occurs can constrain the importance of each mechanism. Fortunately, the well-characterized stars in young and adolescent open clusters (with known ages and composi- tions) provide natural laboratories for such studies, and recent surveys have begun to take advantage

• f this opportunity. We present a review of the discoveries in this emerging realm of exoplanet sci-

ence, discuss the constraints they provide for giant planet formation and migration, and reflect on the future direction of the field.

1 Introduction

Open clusters have long provided key observational constraints in the field of stellar astrophysics due to their unique

• properties. Having formed at the same time (and the same distance), the ensemble properties of the stars in clusters

allow one to determine their ages more precisely than is possible for typical field stars. Moreover, because they formed from the same cloud, the stars in a cluster have approximately the same metal abundance, so differences between stars within a cluster arise primarily as a function of stellar mass. As a result, open clusters enable direct

• bservation of billions of years of stellar evolution (e.g., structure, rotation, activity) across a range of stellar mass.

This allows us to conduct astrophysical experiments under conditions controlled for age, mass, composition, and even a known dynamical environment — such as stellar binary fraction and space density. With the discoveries of thousands of exoplanets over the past two decades1 in a wide range of environments — including hot Jupiters, planets orbiting early-type stars, M dwarfs, giant stars, white dwarfs, in binary systems, and even circumbinary planets — it has become clear that planets are an expected byproduct of star formation. Given that planets are common (e.g., Fressin et al. 2013; Mayor et al. 2011) and most stars form in a clustered environment (e.g., Lada & Lada 2003; Bressert et al. 2010), it is reasonable to expect that planets exist in open

• clusters. This is an exciting prospect, because that means we can begin to use clusters to study planetary evolution

in much the same way that they have been used to study stellar evolution. Planets found in clusters will immediately be among those with the best determined ages, and because open clusters dissipate into the field over time, they tend to be younger than field stars — many have ages < 1 Gyr. It is particularly interesting to note that these ages are similar to the expected timescales for many of the drastic changes that can occur in planetary systems — e.g.,

1http://www.exoplanets.org

29

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold planetary migration, tidal interaction between short-period planets and their host stars, disruptive encounters with passing stars — as well as physical changes to the planets themselves — such as the convergence of cooling tracks for different giant planet and brown dwarf formation models, or atmospheric escape for short-period small planets with gaseous envelopes. We can therefore use the properties of young cluster planets to gain insight into the above processes and the mechanisms that drive them. For example, while hot Jupiters likely formed beyond the ice line and then migrated to within a fraction of an AU of their host stars, the primary channel by which they migrate is unknown. There are many mechanisms that could cause a giant planet to migrate, and two leading ideas are Type II migration (Goldreich & Tremaine 1980, through a gas disk) and high eccentricity migration (HEM) via multi-body interactions, such as planet-planet scattering (Rasio & Ford 1996) or Kozai-Lidov cycles (e.g., Fabrycky & Tremaine 2007). Type II migration must

• ccur while the gas disk is present (∼10 Myr) and is expected to preserve near-circular orbits well-aligned with the

stellar spin axis. Conversely, if HEM is the primary driver of migration, it may take hundreds of millions of years and most hot Jupiters should initially possess non-zero eccentricity and orbital inclination (before subsequent tidal circularization or realignment can occur). Therefore, one way to distinguish between migration mechanisms is to

• bserve systems during, or shortly after, the migration process. If the hot Jupiter occurrence rate is low for stars that

are tens to hundreds of millions of years old, we can rule out disk migration as the dominant mechanism. Similarly, if most young planets have eccentric, inclined orbits, HEM has likely played a large role. Until recently, searches for hot Jupiters in clusters were unsuccessful, despite numerous efforts with sufficient sensitivity, employing both radial velocities (Paulson et al. 2004; Pasquini et al. 2012) and transit photometry (e.g., Bramich et al. 2005; Mochejska et al. 2006, 2008; Rosvick & Robb 2006; Montalto et al. 2007; Pepper et al. 2008; Miller et al. 2008; Hartman et al. 2009). While these results initially seem to be in opposition to the ubiquitous presence of planets, hot Jupiters are relatively more rare (orbiting ∼1% of Sun-like stars; Wright et al. 2012; Mayor et al. 2011), and van Saders & Gaudi (2011) calculate that all combined transit surveys in clusters simply had not surveyed enough stars to guarantee a detection. In this paper, we review the more recent searches for — and discoveries of — planets in clusters and discuss the conclusions that can be drawn from these surveys. We particularly highlight giant planet migration as one topic to which cluster planets can contribute new constraints, and we briefly consider potential future exoplanet studies to be carried out in clusters.

• 50

50 100 150 200 Radial Velocity (m s-1) 0.0 0.2 0.4 0.6 0.8 1.0 Orbital Phase

• 45

45 O-C

• 400
• 200

200 400 Radial Velocity (m s-1) 0.0 0.2 0.4 0.6 0.8 1.0 Orbital Phase

• 55

55 O-C 100 200 300 Radial Velocity (m s-1) 0.0 0.2 0.4 0.6 0.8 1.0 Orbital Phase

• 30

30 O-C

Figure 1: The first two hot Jupiters in an open cluster, Pr0201 b and Pr0211 b (left and middle Quinn et al. 2012), and a Hyades hot Jupiter (right Quinn et al. 2014). The top panels show the radial velocities phased to the orbital periods of 4.4264, 2.1415, and 6.09708 days, and the bottom panels show the residuals to the best fit orbit.

2 Two ‘b’s in the Beehive and the new field of cluster planets

While massive companions had been found previously in open clusters (Sato et al. 2007; Lovis & Mayor 2007), these planets orbit evolved stars of intermediate mass, so they cannot constrain planet occurrence around Sun-like stars in clusters, and their long periods do not provide insight into migration. The discovery of two hot Jupiters in the Praesepe open cluster (also called the Beehive; Quinn et al. 2012, see Fig.1) provided the first evidence that giant planets do form and migrate in dense stellar environments and suggested that they do so at a rate similar to

30

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold that of planets observed around field stars. The subsequent discovery of a hot Jupiter in the Hyades (Quinn et al. 2014) bolstered these conclusions, and allowed a more robust estimate of the hot Jupiter frequency in clusters. By including the null result for the FGK stars in the Hyades surveyed by Paulson et al. (2004), and accounting for completeness of the surveys based on measurement errors and time sampling, Quinn et al. (2014) find that 1.97+1.92

−1.07% of Sun-like stars in the metal-rich, ∼600 Myr Praesepe and Hyades clusters host a hot Jupiter. While this

• ccurrence rate is slightly elevated above the field rate of ∼1%, the errors are still large, and these two clusters are

metal-rich, which is known to correlate with an enhanced frequency of giant planets (e.g., Fischer & Valenti 2005). In fact, when adjusting for this dependence, the cluster hot Jupiter frequency falls to 0.99% at solar metallicity, perfectly consistent with the field. This result is exciting, because it shows that despite the findings of earlier surveys, giant planets not only exist in clusters, but they occur frequently enough for open clusters to be a viable laboratory for the study of the evolution of planetary systems. Table 1: Giant Planets in Open Clusters

Planet Cluster Mp sin i Period Reference (MJup) (days) ǫ Tau b Hyades 7.6 ± 0.2 594.9 ± 5.3 Sato et al. (2007) TYC 5409-2156-1 b NGC2423 10.6 714.3 ± 5.3 Lovis & Mayor (2007) Pr0201 b Praesepe 0.540 ± 0.039 4.4264 ± 0.0070 Quinn et al. (2012) Pr0211 b Praesepe 1.844 ± 0.064 2.1451 ± 0.0012 Quinn et al. (2012) HD 285507 b Hyades 0.917 ± 0.033 6.0881 ± 0.0019 Quinn et al. (2014) YBP1514 b M67 0.40 ± 0.11 5.118 ± 0.001 Brucalassi et al. (2014) YBP1194 b M67 0.34 ± 0.05 6.958 ± 0.001 Brucalassi et al. (2014) SAND364 b M67 1.54 ± 0.24 121.71 ± 0.31 Brucalassi et al. (2014) Pr0211 c Praesepe 7.9 ± 0.2 ∼ > 3500 Malavolta et al. (2016) CB0036 b Coma Ber 2.52 ± 0.19 43.808 ± 0.094 Quinn et al. (2016)

Indeed, the past few years have brought additional discoveries in clusters, with more on the way (see Table 1). Brucalassi et al. (2014) published the results of their 5-year survey in the older M67 cluster, finding two hot Jupiters and one longer period giant planet, the latter of which orbits an evolved star. They find a hot Jupiter frequency in M67 of 2.0+3.0

−1.5%, and the first hint that the frequency of longer period giant planets may also be consistent with

that of the field. More recently Malavolta et al. (2016) presented the discovery of the first multi-planet system in a cluster, with continued monitoring of the Pr0211 hot Jupiter revealing a massive (Mp = 7.9 ± 0.2 MJup) outer companion with a high eccentricity and an orbital period of ∼10 years or more. The orbital architecture of this multi- planet system is suggestive of a dynamical origin for the hot Jupiter (rather than, e.g., gentle Type II migration). Our continued monitoring of the coeval Praesepe, Hyades, and Coma Berenices open clusters has enabled detection of long-period, massive companions in addition to the published hot Jupiters. We detect at least one warm Jupiter (Fig.2) and some warm Jupiter candidates, as well as at least a dozen long-term trends with small amplitudes that are plausibly associated with giant planets or brown dwarfs on long periods (Fig.3). These results in M67, Praesepe, and Coma Berenices highlight the value of long-term monitoring of these cluster stars: understanding the long-period planet population can provide clarity in individual systems (via ob- served orbital architectures), but also provides further context for migration by characterizing the populations of cold Jupiters, which act as a reservoir for the production of hot Jupiters, and warm Jupiters, the properties of which also must fit formation and migration theory. In this way, as a population of cluster planets builds, one can start to consider this class of planet as a whole in the broader context of planet formation and migration. Comparisons against the field, as a function of age, or as a function of the likely birth environment (as characterized by cluster properties) will become more readily investigated, and will contribute much-needed observational constraints to theories of formation and migration.

31

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold

• 200
• 100

100 200 300 Radial Velocity (m s-1) 0.0 0.2 0.4 0.6 0.8 1.0 Orbital Phase

• 100

100 O-C

Figure 2: A warm Jupiter in the Coma Berenices open cluster with a period of 43.808 days and a mass of 2.52 MJup (Quinn et al. 2016).

3 Ensemble evidence for high eccentricity migration

While the population of cluster planets is still small, we can turn to the field population of hot Jupiters for clues about useful lines of investigation. For example, we reiterate here the argument presented in (Quinn et al. 2014) that the dominant migration channel is encoded in the eccentricities hot Jupiters with long tidal circularization

• timescales. The Hyades planet HD 285507 b has a non-zero eccentricity (e = 0.086), which could be indicative of

high eccentricity migration that has been damped to its modest value via tidal interaction with the host star — or it could be the result of another process, such as one-time or ongoing interaction with a third body. In any case, a single planet does not place strong constraints on the typical migration process. Instead, we can examine the eccentricities of field hot Jupiters and compare their estimated ages to the circularization timescale for the system. According to Adams & Laughlin (2006), the circularization timescale is approximately: τcir = 1.6 Gyr × QP 106

• ×

MP MJup

• ×

M∗ M⊙ −1.5 × RP RJup −5 ×

• a

0.05 AU 6.5 (1) The tidal quality factor, QP, is usually assumed to be ∼106 for hot Jupiters, and we can estimate the radii of the planets by using the empirical radius-mass-flux relation of Weiss et al. (2013): RP R⊕ = 2.45 MP M⊕ −0.039 F erg s−1 cm−2 0.094 (2) Systems with short circularization timescales (τcirc < tage) should mostly have circular orbits, and will not help constrain the primary migration mechanism, but those with long circularization timescales (τcirc > tage) — “dynam- ically young” systems — will still retain an imprint of the migration process. We plot these quantities for known hot Jupiters in Fig.4, with dynamically young planets shown to the lower right of the dark solid line and dynami- cally old planets to the upper left. While it is somewhat clear by eye, a Kolmogorov-Smirnov test confirms (with 99.997% confidence) that the eccentricities of dynamically young and dynamically old hot Jupiters are drawn from different populations — dynamically young hot Jupiters are preferentially eccentric. This result supports the idea that hot Jupiters acquire eccentricity during the migration process and then circularize over time, indicating that HEM plays an important role in the production of hot Jupiters. The population of dynamically young hot Jupiters (the lower right in Fig.4) is still small, and finding additional young (cluster) planets is a good way to add to it to strengthen conclusions like this one.

32

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold

• 600
• 400
• 200

Radial Velocity (m/s) 1400 1600 1800 2000 2200 BJD (-2455000)

• 71
• 23

23 71 O-C

• 100

100 200 Radial Velocity (m s-1) 1000 1200 1400 1600 1800 2000 2200 BJD (-2455000)

• 100
• 50

50 100 O-C

Figure 3: A candidate warm Jupiter plus a massive outer comapnion (left), as well as an example of a long-period companion (right). We also note that our result is consistent with investigations of the stellar obliquities for hot Jupiters, which have shown that while many systems are misaligned (indicative of planet-planet scattering or Kozai-Lidov cycles), those with short tidal timescales have orbits well-aligned with the stellar spin (e.g., Albrecht et al. 2012). One of the uncertainties in interpreting these obliquity trends is that a primordially tilted disk would produce inclined orbits regardless of the mode migration; disk migration could still be responsible for hot Jupiter production. The inclusion

• f eccentric dynamically young planets, however, suggests that the inclinations and eccentricities are both excited

by the same process — namely, dynamical interactions with a third body.

4 Future studies of planets in clusters

Armed with the knowledge that planets exist in clusters and can be used to study topics in formation and migration not easily probed by the typically studied planets around field stars, we advocate for continued surveys for planets in clusters, and suggest a few complementary approaches. We also note that within the next couple years, precision distances and kinematics from the Gaia mission will greatly refine and expand membership lists for open clusters, making them even better benchmarks for evolution studies. As described in Sect. 2, continued radial-velocity monitoring of cluster stars can lead to discoveries of not just hot Jupiters, but also long-period planets and the characterization of additional populations that can contribute constraints to migration theory. While the number of very bright nearby clusters that are practically accessible to current spectroscopic observing facilities is somewhat limited, the development of extremely large telescopes will

• pen many more clusters to study. In the meantime, there are still ample targets, particularly in the southern sky,

and in younger clusters. In Sect. 3, we highlighted the value of dynamically young hot Jupiters and suggested that we can add to this population by expanding the search for young planets, though this may require a slight shift in

• methods. In younger clusters (tens to hundreds of millions of years old), rapid rotation and the associated increase

in stellar activity will degrade RV precision (e.g., Lagrange et al. 2013), and the apparent radial-velocity signal induced by a rotating spotted star can mimic that of a short-period planet. Fortunately, the amplitude of this signal is wavelength-dependent (e.g., Reiners et al. 2010), so multi-wavelngth (optical and infrared) spectroscopy has a chance of disentangling stellar activity from bona fide planets. This has been one technique employed by those looking for very young planets around T Tauri stars (e.g., Crockett et al. 2012). While these very young planets are observationally challenging to detect, any hot Jupiter discovered orbiting a T Tauri star would prove that disk migration is effective in some cases. These very young planets can also be searched for via transits, as was done by van Eyken et al. (2012), who reported a candidate hot Jupiter orbiting a T Tauri star. Subsequent studies have

33

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold

• 4
• 2

2 log(τcir) (Gyr)

• 1.0
• 0.5

0.0 0.5 1.0 log(tage) (Gyr)

QP = 2 x 106 QP = 106 QP = 6 x 104 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15 >0.16

Figure 4: From Quinn et al. (2014) — the estimated ages of hot Jupiters plotted against their circularization timescales, and color-coded by their eccentricities (blue points represent circular orbits, and red points represent eccentric orbits). HD 285507 b is indicated by the arrow. The line of equality between τcirc and tage is shown as the dark, solid diagonal line, with dynamically young planets to the lower right and dynamically old planets to the upper left. A K-S test confirms a difference in the populations with 99.997% confidence. called its validity into question, but T Tauri stars remain an interesting targets for planet searches. As we find evidence for long period companions in clusters (Fig.3), these objects can be followed up with high- resolution imaging such as non-redundant aperture masking interferometry from Keck, which would be sensitive to massive brown dwarfs beyond about 25 AU, and objects as small as ∼20 MJup at 100 AU in the Coma Berenices cluster (see Fig.5). While the imaging would not be sensitive to planetary mass companions, both null results and detections would better characterize the architectures of planetary systems. Moreover, wide companions in planetary systems in clusters are of particular interest, as some fraction of these may be disrupted by the time they are observed as field stars. Finally, characterizing the populations of both planets and directly imaged brown dwarfs at ages of several hundred million years would provide a link between the typically young populations studied by direct imaging surveys and the older stars typically studied by radial velocity and transit surveys. While this conference has focused on giant planets, we would be remiss to not acknowledge the upcoming promise of space-based transit surveys, such as NASA’s K2 and TESS missions in the context of cluster planets. While most planets found by these missions will be small, they will still hold valuable clues for planetary evolution, and because they transit, it also becomes possible to study the evolution of their interior structures and atmospheres. The Kepler open cluster study (Meibom et al. 2013) found two mini-Neptunes in a cluster, and calculated that small planets are also about as abundant in clusters as in the field, so we can expect more cluster planets from K2 and TESS, which will observe many more open clusters than did the original Kepler mission. The first of these discoveries, a small planet in the Hyades (Mann et al. 2015), was recently announced. Such transiting cluster planets — with known ages, compositions, radii, and (hopefully) masses — will be benchmark objects upon which we base our understanding of the evolution of planetary atmospheres and interior structures.

5 Conclusion

Until four years ago, it was unclear whether short period planets existed in clusters, but several recent surveys have shown that they occur at rates similar to that of isolated stars like the Sun. This opens up new fields of study that

34

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold

1 10 100 Separation (AU) 0.01 0.10 1.00 Companion mass (MSun) Coma Ber

RV AO+NRM

1 10 100 Separation (AU) 0.01 0.10 1.00 Companion mass (MSun)

Figure 5: Approximate detection limits for our existing radial velocities (blue) and typical non-redundant aperture masking interferometry with Keck (red) in the Coma Berenices open cluster. The combination would be sensitive to nearly all stellar companions, and many brown dwarfs — at all separations. make use of the unique properties of open clusters. Using the known ages, compositions, and dynamical environ- ments of clusters, we can study the evolution of a wide range of planetary properties, including their orbits, system architectures, physical structures, and atmospheres, and we can do so while controlling for potential obfuscating properties like composition and birth environment. While interesting individual discoveries and useful constraints for migration theory have already emerged from this fledgling field of study, the future of open cluster exoplanet science promises to deliver many more discoveries (including small planets), and should make vital contributions to physical and dynamical evolutionary studies of exoplanetary systems. Acknowledgments: SQ is supported by the NSF Graduate Research Fellowship, Grant DGE-1051030. We thank Trent Dupuy for calculating imaging detection limits and producing the associated figure. We also thank the con- ference organizers for a wonderful conference. We are all eagerly anticipating “One Hundred Years of Giant Exoplanets” in 2095.

References

Adams, F. C. & Laughlin, G. 2006, ApJ, 649, 1004 Albrecht, S., Winn, J. N., Johnson, J. A., et al. 2012, ApJ, 757, 18 Bramich, D. M., Horne, K., Bond, I. A., et al. 2005, MNRAS, 359, 1096 Bressert, E., Bastian, N., Gutermuth, R., et al. 2010, MNRAS, 409, L54 Brucalassi, A., Pasquini, L., Saglia, R., et al. 2014, A&A, 561, L9 Crockett, C. J., Mahmud, N. I., Prato, L., et al. 2012, ApJ, 761, 164 Fabrycky, D. & Tremaine, S. 2007, ApJ, 669, 1298 Fischer, D. A. & Valenti, J. 2005, ApJ, 622, 1102 Fressin, F., Torres, G., Charbonneau, D., et al. 2013, ApJ, 766, 81

35

Twenty years of giant exoplanets - Proceedings of the Haute Provence Observatory Colloquium, 5-9 October 2015 Edited by I. Boisse, O. Demangeon, F. Bouchy & L. Arnold Goldreich, P. & Tremaine, S. 1980, ApJ, 241, 425 Hartman, J. D., Gaudi, B. S., Holman, M. J., et al. 2009, ApJ, 695, 336 Lada, C. J. & Lada, E. A. 2003, ARAA, 41, 57 Lagrange, A.-M., Meunier, N., Chauvin, G., et al. 2013, A&A, 559, A83 Lovis, C. & Mayor, M. 2007, A&A, 472, 657 Malavolta, L., Nascimbeni, V., Piotto, G., et al. 2016, A&A, accepted Mann, A. W., Gaidos, E., Mace, G. N., et al. 2015, ArXiv e-prints Mayor, M., Marmier, M., Lovis, C., et al. 2011, ArXiv e-prints Meibom, S., Torres, G., Fressin, F., et al. 2013, Natur, 499, 55 Miller, A. A., Irwin, J., Aigrain, S., Hodgkin, S., & Hebb, L. 2008, MNRAS, 387, 349 Mochejska, B. J., Stanek, K. Z., Sasselov, D. D., et al. 2006, AJ, 131, 1090 Mochejska, B. J., Stanek, K. Z., Sasselov, D. D., et al. 2008, ACA, 58, 263 Montalto, M., Piotto, G., Desidera, S., et al. 2007, A&A, 470, 1137 Pasquini, L., Brucalassi, A., Ruiz, M. T., et al. 2012, A&A, 545, A139 Paulson, D. B., Cochran, W. D., & Hatzes, A. P. 2004, AJ, 127, 3579 Pepper, J., Stanek, K. Z., Pogge, R. W., et al. 2008, AJ, 135, 907 Quinn, S. N., White, R. J., Latham, D. W., et al. 2012, ApJL, 756, L33 Quinn, S. N., White, R. J., Latham, D. W., et al. 2016, in prep Quinn, S. N., White, R. J., Latham, D. W., et al. 2014, ApJ, 787, 27 Rasio, F. A. & Ford, E. B. 1996, Science, 274, 954 Reiners, A., Bean, J. L., Huber, K. F., et al. 2010, ApJ, 710, 432 Rosvick, J. M. & Robb, R. 2006, AJ, 132, 2309 Sato, B., Izumiura, H., Toyota, E., et al. 2007, ApJ, 661, 527 van Eyken, J. C., Ciardi, D. R., von Braun, K., et al. 2012, ApJ, 755, 42 van Saders, J. L. & Gaudi, B. S. 2011, ApJ, 729, 63 Weiss, L. M., Marcy, G. W., Rowe, J. F., et al. 2013, ApJ, 768, 14 Wright, J. T., Marcy, G. W., Howard, A. W., et al. 2012, ApJ, 753, 160

36