Speckle Observations of TESS Exoplanet Host Stars. II. Stellar Companions at 1-1000 AU and Implications for Small Planet Detection
Total Page：16
File Type：pdf, Size：1020Kb
Draft version June 28, 2021 Typeset using LATEX twocolumn style in AASTeX63 Speckle Observations of TESS Exoplanet Host Stars. II. Stellar Companions at 1-1000 AU and Implications for Small Planet Detection Kathryn V. Lester,1 Rachel A. Matson,2 Steve B. Howell,1 Elise Furlan,3 Crystal L. Gnilka,1 Nicholas J. Scott,1 David R. Ciardi,3 Mark E. Everett,4 Zachary D. Hartman,5, 6 and Lea A. Hirsch7 1NASA Ames Research Center, Moffett Field, CA 94035, USA 2U.S. Naval Observatory, Washington, D.C. 20392, USA 3NASA Exoplanet Science Institute, Caltech/IPAC, Pasadena, CA 91125, USA 4NSF's National Optical-Infrared Astronomy Research Laboratory, Tucson, AZ 85719, USA 5Lowell Observatory, Flagstaff, AZ 86001, USA 6Department of Physics & Astronomy, Georgia State University, Atlanta GA 30303, USA 7Kavli Institute for Particle Astrophysics and Cosmology, Stanford University, Stanford, CA 94305, USA (Accepted June 17, 2021) ABSTRACT We present high angular resolution imaging observations of 517 host stars of TESS exoplanet candi- dates using the `Alopeke and Zorro speckle cameras at Gemini North and South. The sample consists mainly of bright F, G, K stars at distances of less than 500 pc. Our speckle observations span angular resolutions of ∼20 mas out to 1.2 arcsec, yielding spatial resolutions of <10 to 500 AU for most stars, and our contrast limits can detect companion stars 5−9 magnitudes fainter than the primary at optical wavelengths. We detect 102 close stellar companions and determine the separation, magnitude differ- ence, mass ratio, and estimated orbital period for each system. Our observations of exoplanet host star binaries reveal that they have wider separations than field binaries, with a mean orbital semi-major axis near 100 AU. Other imaging studies have suggested this dearth of very closely separated binaries in systems which host exoplanets, but incompleteness at small separations makes it difficult to disen- tangle unobserved companions from a true lack of companions. With our improved angular resolution and sensitivity, we confirm that this lack of close exoplanet host binaries is indeed real. We also search for a correlation between planetary orbital radii vs. binary star separation, but given the very short orbital periods of the TESS planets, we do not find any clear trend. We do note that in exoplanet systems containing binary host stars, there is an observational bias against detecting Earth-size planet transits due to transit depth dilution caused by the companion star. Keywords: binary stars, exoplanets, high angular resolution, speckle interferometry 1. INTRODUCTION of 4" for Kepler/K2 and 21" for TESS, nearby stars or The Kepler, K2, and TESS missions have provided unresolved stellar companions within the aperture dilute precise, light curve photometry for millions of stars and the transit depth and bias the measured planet radius arXiv:2106.13354v1 [astro-ph.SR] 24 Jun 2021 contributed breakthrough advancements in many fields (Ciardi et al. 2015). About 50% of solar-type exoplanet of stellar astrophysics (Borucki et al. 2010; Howell et al. hosts have companions (e.g., Horch et al. 2014; Mat- 2014; Ricker et al. 2015). This includes the discovery son et al. 2018), so follow-up high resolution imaging of thousands of extra-solar planets through the transit is needed to search for unresolved companions and aid method, where a planet passing in front of a star causes exoplanet validation and characterization. a periodic dip in brightness. Due to the large pixel sizes Understanding how binary companions affect the for- mation, evolution, and survival of exoplanets is also an important component to understanding planet forma- Corresponding author: Kathryn Lester tion overall. Theoretical studies show that a compan- [email protected] ion can truncate the protoplanetary disk, leaving less 2 material for planets to form (Martin et al. 2014; Jang- for fainter targets (V > 9 mag), and a point source stan- Condell 2015), cause the migration of gas giant planets dard star was observed immediately before or after each (Dawson & Johnson 2018), or disperse the disk before target for calibration. planets are able to form (Cieza et al. 2009; Kraus et al. The data were reduced using the pipeline developed 2012). Observational studies using high resolution imag- by the speckle team (Howell et al. 2011; Horch et al. ing (Kraus et al. 2016; Fontanive et al. 2019; Ziegler et 2011) to calculate and average the power spectrum of al. 2020, 2021; Howell et al. 2021) and radial velocities each image, then correct for the speckle transfer function (Wang et al. 2014; Hirsch et al. 2021) have found fewer by dividing the mean power spectrum of the target by companions within 100 AU around exoplanet hosts than that of the standard star. The pipeline also produces a around field stars, supporting the idea that close stellar reconstructed image of each target with a field of view of companions do suppress planet formation. 1.3" in radius around the target. The contrast limits for However, it is difficult to disentangle unobserved com- each target were determined with the method described panions from a true lack of companions without suf- in Horch et al.(2011), which uses the background flux ficient angular resolution. The angular separations of levels to determine the faintest companions one could the closest companions (< 50 AU) around Kepler exo- reliably detect at each separation. These 5σ contrast planet hosts are near or below the diffraction limit of limits at 0.200 and 1.000 in the blue and red filters are most telescopes, but TESS observed brighter, nearby also listed in Tables3 and4. stars for which we can resolve companions at smaller Due to the large TESS pixels, some transit detections physical separations from the host star. We present the were identified to be false positives through follow-up highest angular resolution speckle images of TESS exo- photometry. If the light curve shows chromaticity or planet host stars in search of companions at 1 − 100 AU v-shaped dips of alternating depth, the event must be where planet formation would be greatly hindered. We a binary eclipse rather than a planet transit, then con- describe our observations and companion detections in firmed with follow-up observations to identify the eclips- Section2, then create a simulated sample of binary stars ing source within the TESS aperture. The transits for to investigate the separation distribution of exoplanet 105 of the TOI's in our sample have been identified as host binaries in Section3. Our results and conclusions false positives (or false alarms) on the Exoplanet Follow- are presented in Sections4 and5, respectively. up Observing Program (ExoFOP)2 website, noted in Ta- bles3 and4. We have eliminated these host stars from 2. OBSERVATIONS further consideration here, leaving us with 412 stars. 2.1. Speckle Imaging A few additional systems were classified as ambiguous planetary candidates (APC), for example, if only a sin- We observed 517 stars from the TESS Objects of Inter- gle transit was observed. Because these systems have est (TOI) catalog using the `Alopeke and Zorro speckle not yet been confirmed as false positives, we kept them cameras (Scott et al. 2018, 2021) on the Gemini 8.1 m in our sample and discuss their impact in Section 4.1. North and South telescopes respectively from May 2019 Figure1 shows the distributions in effective tem- to December 2020. Before each observing season, tar- perature, distance, and TESS magnitude for our 412 gets were selected from the latest version of the TOI exoplanet host TOI's. Our sample contains primar- catalog of likely planet candidates. All of our obser- ily bright, Solar-type stars with smaller distances from vations are listed in Tables3 and4 in the Appendix, Earth compared to Kepler host stars, allowing Gemini with the TOI and TIC numbers, TESS magnitude, UT speckle observations to detect companions quite close to date, the inverse of the parallax from Gaia EDR3 (Gaia the target star. For distances of 100 − 500 pc, our in- Collaboration et al. 2016, 2020) for the distance, effec- ner angular resolution of 0.017" at 562 nm corresponds tive temperature of the host star from the TESS Input to separations of 1:7 − 8:5 AU, while our inner angular Catalog (TIC v8.1, Stassun et al. 2019), and notes for resolution of 0.026" at 832 nm corresponds to separa- each target. The Julian date of each observation can be tions of 2:6 − 13:0 AU. Therefore, we can successfully found in the headers of the archival data hosted on the detect stellar companions in the 1 − 100 AU regime and 1 Gemini Observatory Archive . At least three image sets determine their properties. were obtained for each target, where one set consists of 1000 60 ms exposures taken in a 562 nm and an 832 nm filter simultaneously. Additional image sets were taken 2 https://exofop.ipac.caltech.edu/tess 1 https://archive.gemini.edu/searchform 3 40 30 20 # of TOIs 10 0 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 Effective Temperature (K) 80 60 40 # of TOIs 20 0 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Distance (pc) 30 20 # of TOIs 10 0 6 7 8 9 10 11 12 13 14 15 16 17 TESS Magnitude (mag) Figure 1. Stellar parameter distributions for the 412 exoplanet host TOI's in our sample, including stellar effective temperature (top), Gaia EDR3 distance (middle), and TESS magnitude (bottom). 2.2. Detected Companions in the 832 nm filter, and could not constrain the magni- Binary stars produce a characteristic interference tude difference for three of these systems. Results from fringe pattern in the power spectrum from which the multiple nights were averaged together because no or- companion's angular separation (ρ), position angle (θ, bital motion could be seen, so any TOI's with dupli- measured East of North), and magnitude difference cate entries in Table1 were found to be triple systems.