UC Berkeley

Nearly half of all solar-type stars have at least one stellar or brown dwarf companion, and planets around G- and K-type stars appear to be quite common. Yet the impact of stellar multiplicity on planets, and on our understanding of planets, is not well understood. In this thesis, I describe a combination of work to study the interplay between stellar and planetary companions, using the combination of radial velocity and high-resolution imaging observations.

First, planets discovered by transit surveys like the \eke Mission are sensitive to a detection and characterization bias due to stellar companions. The light of unresolved binary companions dilutes the transit of the planet, making it appear shallower than it is. This causes planet sizes in binary systems to be underestimated, and sensitivity to small planets to be overestimated. In Chapter 2, I combined multi-wavelength observations of stellar companions detected within 2 arcseconds of 170 Kepler Object of Interest hosts, to assess whether these companions were physically associated with the planet host stars. I found that stellar companions within 1" had a >80% likelihood of being bound. I used the compiled imaging observations of the bound stellar companions in this sample to determine the factor by which planet radii in these systems were underestimated. I determined that an average radius correction factor of X_R = 1.65 must be applied to planets orbiting stars with unresolved stellar companions, assuming the planets are equally likely to orbit either of the stars in a binary system. This means that planets in unresolved binary systems may have radii underestimated by an average of 65%. This effect is particularly important for Kepler stars with minimal ground-based follow-up; by performing a combination of spectroscopic and imaging observations on Kepler planet hosts, it is possible to reduce the effect of stellar companions by finding or ruling out binary companions in large regions of parameter space.

Second, stellar companions likely impact the formation and evolution of planetary systems. To probe these effects, I focused on the solar neighborhood, studying sun-like stars within 25 pc using Doppler spectroscopy as well as adaptive optics and speckle imaging techniques. I built up methodology for performing joint orbital fits to combinations of radial velocity, astrometric, and photometric data by assessing the orbit of HD 159062 in Chapter 3. HD 159062 is a nearby binary system consisting of a G dwarf and a newly-discovered white dwarf companion. I determined that the white dwarf must have a mass of 0.54^{+0.09}_{-0.03} solar masses. Its combined radial velocities, astrometry, and photometry are inconsistent with it being either a brown dwarf or a main sequence star.

In Chapter 4, I extended this joint orbital analysis technique to make use of imaging non-detections to place constraints on orbiting planets. In this study, we focused on the nearby adolescent K dwarf, Epsilon Eridani, whose giant planet was detected in three decades of radial velocity monitoring. Using deep M-band coronographic imaging that resulted in no detection of the planet, upper limits were placed on the mass of the giant planet. The combination of the RV data and constraints from the non-detection of the planet allowed the mass and inclination of the planet to be determined more precisely than with the RV data alone.

Finally, I performed a simultaneous RV and imaging survey for both stellar and planetary companions to sun-like stars in the solar neighborhood in Chapter 5. By comparing the planet occurrence rates in binary systems to those in single star systems, we can begin to untangle the effects of stellar companions on planet formation. I found that the occurrence rate of planets with masses between 0.1 and 10 Jupiter masses and with semi-major axes between 0.1 and 10 AU is statistically equivalent in single star and binary star systems. I also determined that all known planet-hosting binary systems in the 25 pc sun-like star sample had separations wider than 100 AU. This indicates that wide binary companions do not seem to strongly impact the planet formation process, but that the planet occurrence rate in binary systems may decrease for systems with smaller binary separations.