The Gemini Deep Planet Survey -- GDPS
David Lafreniere, Rene Doyon, Christian Marois, Daniel Nadeau, Ben R. Oppenheimer, Patrick F. Roche, Francois Rigaut, James R. Graham, Ray Jayawardhana, Doug Johnstone, Paul G. Kalas, Bruce Macintosh, Rene Racine
aa r X i v : . [ a s t r o - ph ] A ug D RAFT VERSION N OVEMBER
11, 2018
Preprint typeset using L A TEX style emulateapj v. 08/13/06
THE GEMINI DEEP PLANET SURVEY – GDPS ∗ D AVID L AFRENIÈRE A , R ENÉ D OYON A , C HRISTIAN M AROIS B , D ANIEL N ADEAU A , B EN R. O
PPENHEIMER C , P ATRICK
F. R
OCHE D ,F RANÇOIS R IGAUT E , J AMES
R. G
RAHAM F , R AY J AYAWARDHANA G , D OUG J OHNSTONE H , P AUL
G. K
ALAS F , B RUCE M ACINTOSH B ,R ENÉ R ACINE A Draft version November 11, 2018
ABSTRACTWe present the results of the Gemini Deep Planet Survey, a near-infrared adaptive optics search for giantplanets and brown dwarfs around nearby young stars. The observations were obtained with the Altair adaptiveoptics system at the Gemini North telescope and angular differential imaging was used to suppress the specklenoise of the central star. Detection limits for the 85 stars observed are presented, along with a list of all faintpoint sources detected around them. Typically, the observations are sensitive to angular separations beyond0.5 ′′ with 5 σ contrast sensitivities in magnitude difference at 1.6 µ m of 9.5 at 0.5 ′′ , 12.9 at 1 ′′ , 15.0 at 2 ′′ ,and 16.5 at 5 ′′ . For the typical target of the survey, a 100 Myr old K0 star located 22 pc from the Sun, theobservations are sensitive enough to detect planets more massive than 2 M Jup with a projected separation in therange 40–200 AU. Depending on the age, spectral type, and distance of the target stars, the detection limit canbe as low as ∼ M Jup . Second epoch observations of 48 stars with candidates (out of 54) have confirmed that allcandidates are unrelated background stars. A detailed statistical analysis of the survey results, yielding upperlimits on the fractions of stars with giant planet or low mass brown dwarf companions, is presented. Assuminga planet mass distribution d n / d m ∝ m - . and a semi-major axis distribution d n / d a ∝ a - , the 95% credibleupper limits on the fraction of stars with at least one planet of mass 0.5–13 M Jup are 0.28 for the range 10–25 AU, 0.13 for 25–50 AU, and 0.093 for 50–250 AU; this result is weakly dependent on the semi-major axisdistribution power-law index. The 95% credible interval for the fraction of stars with at least one brown dwarfcompanion having a semi-major axis in the range 25–250 AU is 0 . + . - . , irrespective of any assumption onthe mass and semi-major axis distributions. The observations made as part of this survey have resolved thestars HD 14802, HD 166181, and HD 213845 into binaries for the first time. Subject headings:
Planetary systems — stars: imaging — binaries: close — stars: low-mass, brown dwarfs INTRODUCTIONMore than 200 exoplanets have been discovered over thelast decade through precise measurements of variations of theradial velocity (RV) of their primary star. Besides establish-ing that at least 6–7% of FGK stars have at least one giantplanet with a semi-major axis smaller than ∼ ∗ Based on observations obtained at the Gemini Observatory, which is op-erated by the Association of Universities for Research in Astronomy, Inc.,under a cooperative agreement with the NSF on behalf of the Gemini partner-ship: the National Science Foundation (United States), the Particle Physicsand Astronomy Research Council (United Kingdom), the National ResearchCouncil (Canada), CONICYT (Chile), the Australian Research Council (Aus-tralia), CNPq (Brazil) and CONICET (Argentina).Electronic address: [email protected] [email protected] A Département de physique and Observatoire du Mont Mégantic, Univer-sité de Montréal, C.P. 6128 Succ. Centre-Ville, Montréal, QC, H3C 3J7,Canada B Institute of Geophysics and Planetary Physics L-413, Lawrence Liver-more National Laboratory, 7000 East Ave, Livermore, CA 94550 C Department of Astrophysics, American Museum of Natural History,Central Park West at 79th Street, New York, NY 10024 D Astrophysics, Physics Department, University of Oxford, 1 Keble Road,Oxford, OX1 3RH, UK E Gemini Observatory, Southern Operations Center, Association of Uni-versities for Research in Astronomy, Inc., Casilla 603, La Serena, Chile F Department of Astronomy, University of California at Berkeley, 601Campbell Hall, Berkeley, CA 94720 G Department of Astronomy and Astrophysics, University of Toronto, 50St. George Street, Toronto, ON, M5S 3H4, Canada H National Research Council Canada, Herzberg Institute of Astrophysics,5071 West Saanich Road, Victoria, BC, V9E 2E7, Canada reader is referred to Udry et al. (2007); Butler et al. (2006);Marcy et al. (2005). Besides the RV technique, the photo-metric transit method has lead successfully to the discoveryof new exoplanets on small orbits (e.g. Konacki et al. 2003;Alonso et al. 2004; Cameron et al. 2007) and has provided thefirst measurements of the radius and mean density of giant ex-oplanets (e.g. Charbonneau et al. 2000). Very recently, a fewexoplanets have been detected by gravitational microlensing(Bond et al. 2004; Udalski et al. 2005; Beaulieu et al. 2006;Gould et al. 2006); these planets have separations of ∼ ∼
10 AU. As aresult, the population of exoplanets on large orbits is currentlyunconstrained.The two main models of giant planet formation are coreaccretion (Pollack et al. 1996) and gravitational instability(Boss 1997, 2001). In the core accretion model, solid particleswithin a proto-planetary disk collide and grow into solid coreswhich, if they become massive enough before the gas disk dis-sipates, trigger runaway gas accretion and become giant plan-ets. Models predict that the timescale for formation of a planetlike Jupiter through this process is about 5 Myr (Pollack et al.1996), or about 1 Myr if migration of the core through the diskis allowed as the planet forms (Alibert et al. 2005). Thesetimescales are comparable to or below the estimated proto-planetary dust disk lifetime ( ∼ .
10 Myr, Jayawardhana et al. 2006). For-mation through core accretion is strongly dependent on thesurface density of solid material (hence [Fe/H]) in the proto-planetary disk , precluding formation of Jupiter mass plan- Lafrenière et al.ets at distances greater than 15–20 AU (e.g. Pollack et al.1996; Ida & Lin 2004), where the low density of planetesi-mals would lead to prohibitively long formation timescales.Neptune mass planets can be formed out to slightly larger dis-tances and can further migrate outward owing to interactionwith the disk.In the gravitational instability model, small instabilities in aproto-planetary disk grow rapidly into regions of higher den-sity that subsequently evolve into spiral arms owing to Ke-plerian rotation. Further interactions between these spiralarms lead to the formation of hot spots which then collapseto form giant planets. The range of orbital separation overwhich this mechanism may operate efficiently is not yet clear.Some studies indicate that it may lead to planet formation onlyat separations exceeding ∼
100 AU (Whitworth & Stamatel-los 2006; Matzner & Levin 2005), where the radiative cool-ing timescale is sufficiently short compared to the dynamicaltimescale, while others have been able to produce planets onlyat separations below 20–30 AU (Boss 2000, 2003, 2006).A few other models are capable of forming giant planets onlarge orbits directly. One such mechanism is shock-inducedformation following collision between disks (Shen & Wad-sley 2006). In this model, the violent collision of two proto-planetary disks triggers instabilities that lead to the collapse ofplanetary or brown dwarf (BD) mass clumps. Results of nu-merical simulations indicate that planets and BDs may format separations of several tens of AU or more through this pro-cess (Shen & Wadsley 2006). The competitive accretion andejection mechanism that was proposed initially to explain theformation of BDs (Reipurth & Clarke 2001) could also formplanetary mass companions on large orbits, as suggested bythe results of recent simulations by Bate & Bonnell (2005).Even in a scenario in which all giant planets form on smallorbits, through either core accretion or gravitational collapse,a significant fraction of planets could be found on stable or-bits of tens of AU because of outward orbital migration. In-deed, numerical simulations have shown that gravitational in-teractions between planets in a multi-planet system may sendone of the planets, usually the least massive one, out to aneccentric orbit of semi-major axis of tens to hundreds of AU(Chatterjee et al. 2007; Veras & Armitage 2004; Rasio & Ford1996; Weidenschilling & Marzari 1996). This process couldbe involved frequently in the shaping of the orbital parame-ters of planetary systems as we have learned from RV surveysthat multi-planet systems are common, representing ∼
14% ofknown planetary systems (Marcy et al. 2005). Outward mi-gration of massive planets can be induced also by interactionsbetween the planet and the gaseous disk; the simulations ofVeras & Armitage (2004) reveal that this process is capable ofcarrying Jupiter mass planets out to several tens of AU. Sim-ilarly, angular momentum exchange between two planets (ormore), achieved through viscous interactions with the disk,could drive the outer planet to a separation of hundreds of AU(Martin et al. 2007). Outward planet migration can result fur-ther from interaction of the planet with the solid particles inthe disk after the gas has dissipated (e.g. Levison et al. 2007);there is in fact strong evidence that this mechanism has playedan important role in the Solar system (Fernandez & Ip 1984;Malhotra 1995; Hahn & Malhotra 2005). Based on numericalsimulations, it is likely that all giant planets of the Solar sys-tem formed interior to ∼
15 AU and migrated outward (exceptJupiter) to their current location (Tsiganis et al. 2005).From an observational point of view, there is some evidencethat planets on large orbits may exist. Many observations of dusty disks around young stars, made either in emitted light(e.g. Vega, ε Eri, Fomalhaut; Holland et al. 1998; Greaveset al. 1998) or in scattered light (e.g. HD 141569, HR 4796,Fomalhaut; Augereau et al. 1999; Weinberger et al. 1999;Schneider et al. 1999; Kalas et al. 2005), have unveiled asym-metric or ring-like dust distributions. These peculiar mor-phologies could arise from gravitational dust confinement im-posed by one or more (unseen) giant planets on orbits of tensto hundreds of AU. In fact, detailed numerical simulations ofthe effect of giant planets on the dynamical evolution of dustydisks have been able to reproduce the observed morphologieswith remarkable agreement (Ozernoy et al. 2000; Wilner et al.2002; Deller & Maddison 2005). Typically, Jupiter mass plan-ets on orbits of ∼
60 AU are needed to reproduce the observa-tions, although in some cases less massive planets (similar toNeptune) may be able to reproduce the observed features.In the last few years, there have been a few discoveriesof planetary mass or low-mass BD companions located be-yond several tens of AU, in projection, from their primary: an ∼ M Jup companion 40 AU from the BD 2M 1207 - ∼ M Jup companion 100 AU from the T Tauri star GQ Lup (Maroiset al. 2007a; Seifahrt et al. 2007; Neuhäuser et al. 2005), a ∼ M Jup companion 210 AU from the young star CHXR 73(Luhman et al. 2006), a ∼ M Jup companion 240 AU fromthe young BD 2M 1101 - ∼ M Jup companion 260 AU from the young star AB Pic (Mohantyet al. 2007; Chauvin et al. 2005b), a 7–19 M Jup companion240–300 AU from the young BD Oph 1622 - ∼ M Jup companion 330 AU from the T Tauri star DH Tau(Luhman et al. 2006; Itoh et al. 2005), and a ∼ M Jup com-panion 790 AU from the star HN Peg (Luhman et al. 2007b).These discoveries might indicate that more similar compan-ions, and less massive ones, do exist and remain to be found.Perhaps even more compelling is the fact that the numberof exoplanets found by RV surveys increases as a function ofsemi-major axis for the range 0.1–3 AU (Butler et al. 2006);these surveys are incomplete at larger separations. Conserva-tive extrapolation suggests that there may be at least as manyplanets beyond 3 AU as there are within (Butler et al. 2006).In fact, long-term trends in RV data have been detected forabout 5% of the stars surveyed (Marcy et al. 2005), suggest-ing the presence of planets between 5 AU and 20 AU aroundthem.Given all of the considerations above, it is clear that a de-termination of the frequency of giant planets as a function oforbital separation out to hundreds of AU is necessary to elu-cidate the relative importance of the various modes of planetformation and migration. Direct imaging is currently the onlyviable technique to probe for planets on large separations andachieve this goal. However, detecting giant planets directlythrough imaging is very difficult due to the angular proxim-ity of the star and the very large luminosity ratios involved.Currently, the main technical difficulty when trying to imagegiant planets directly does not come from diffraction of lightby the telescope aperture, from light scattering due to resid-ual atmospheric wavefront errors after adaptive optics (AO)correction, nor from photon noise of the stellar point spreadfunction (PSF), but rather from light scattering by optical im-perfections of the telescope and camera that produce brightquasi-static speckles in the PSF of the central star. Thesespeckles are usually much brighter than the planets soughtafter. More in depth discussions of this problem, as well ashe Gemini Deep Planet Survey 3 N u m be r o f s t a r s F0 F5 G0 G5 K0 K5 M0 M5Spectral type051015202530 N u m be r o f s t a r s
10 100 1000 10000Age (Myr)05101520 N u m be r o f s t a r s F IG . 1.— Distribution of distance, spectral type, and age of the target stars. For the age distribution, each star was distributed over all the age bins according tothe fraction of their estimated age interval falling inside each bin. possible venues to circumvent it using current instrumenta-tion, can be found in Lafrenière et al. (2007); Hinkley et al.(2007); Marois et al. (2006, 2005); Masciadri et al. (2005);Biller et al. (2004); Schneider & Silverstone (2003); Maroiset al. (2003); Sparks & Ford (2002); Marois et al. (2000);Racine et al. (1999). As AO systems continue to improveand eventually achieve Strehl ratios above ∼ J , H , or K ,or Kasper et al. (2007) and Heinze et al. (2006) for searchesmade in L ′ or M ′ . Depending on the observing strategy em-ployed, the properties of the target stars, and the characteris-tics of the instrument used, each of these surveys was sensitiveto a different regime of companion masses and separations.Typically, these surveys have reached detection contrasts of10–13 mag for angular separations beyond 1 ′′ –2 ′′ , sufficientto detect planets more massive than ∼ M Jup for targets aged ∼
100 Myr. Unfortunately, rigorous statistical analyses allow-ing derivation of clear constraints on the population of planetsin the regimes of mass and separation to which these surveyswere sensitive are only beginning to be reported in the litera-ture ; an assessment of the current status of knowledge is thusrather difficult to make. Nonetheless, it is fair to say that thepopulation of planets less massive than ∼ M Jup , having orbitswith a semi-major axis of tens to hundreds of AU, is poorlyconstrained.In this paper we report the results of the Gemini DeepPlanet Survey (GDPS), a direct imaging survey of 85 nearbyyoung stars aimed at constraining the population of Jupitermass planets with orbits of semi-major axis in the range 10-300 AU. The selection of the GDPS target sample is explainedin §2, and the observations and data reduction are detailed in§3. The detection limits achieved for each target are then pre-sented in §4 along with all candidate companions detected. A In addition to the present work, analyses by Nielsen et al. (2007) andKasper et al. (2007) have become available during the review process of thismanuscript. statistical analysis of the results allowing determination of themaximum fraction of stars that could bear planetary compan-ions is presented in §5. Concluding remarks follow in §6. TARGET SAMPLEIn light of the luminosity ratio and angular separation prob-lem highlighted above, the list of target stars was assembledmainly on the basis of young age and proximity to the Sun,the latter yielding a larger angular separation for a given phys-ical distance between the star and an eventual planet. Equiv-alently, a given detection threshold is achieved at a smallerphysical separation for a star closer to the Sun, and planets onsmaller orbits can be detected. Additionally, for angular sep-arations where planet detection is limited by sky backgroundnoise or read noise, lower mass planets can be detected arounda star closer to the Sun as their apparent brightness would belarger. Giant planets are intrinsically more luminous at youngages and fade with time (e.g. Marley et al. 2007; Baraffe et al.2003; Burrows et al. 1997); therefore, for a given detectionthreshold, observations of younger stars are sensitive to plan-ets having a lower mass. The proximity and age criteria usedin building the target list thus maximize the range of mass andseparation over which the survey is sensitive.The target stars were selected from three sources: (1) Ta-bles 3 and 4 of Wichmann et al. (2003), which list nearbystars with an estimated age below or comparable to that ofthe Pleiades ( ∼
100 Myr), based on measurements of lithiumabundance, space velocity, and X-ray activity; (2) Tables 2and 5 of Zuckerman & Song (2004b), which list members ofthe β Pictoris ( ∼
12 Myr) and AB Doradus ( ∼
50 Myr) mov-ing groups respectively; and (3) Tables 2 and 5 of Monteset al. (2001b), which list late-type single stars that are possiblemembers of the Local Association (Pleiades moving group,20–150 Myr) and IC 2391 supercluster (35–55 Myr) respec-tively, based on space velocity measurements. The stars listedin Montes et al. (2001b) were initially selected based on vari-ous criteria indicative of youth, such as kinematic properties,rotation rate, chromospheric activity, lithium abundance, orX-ray emission, but for many of these stars the space velocityis the only indication of youth as other measurements are ei-ther unavailable or inconclusive; the young age of such starsis therefore uncertain. This uncertainty will be taken into ac-count in our statistical analysis (§5). A few stars known tohave a circumstellar disk were added to these lists. L a fr e n i è r ee t a l . TABLE 1GDPS
TARGET SAMPLE
Names α δ
Spectral H Dist. a µ α cos δ a µ δ a [Fe/H] [Fe/H] Age Age NotesHD GJ HIP Other (J2000) (J2000) Type (mag) (pc) (mas/yr) (mas/yr) Ref. (Myr) Ref.166 5 544 - 00 h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . . h m . s + ◦ ′ . ′′ b . b - . b -0.78 N04 50–50 Z04b AB Dor5996 - 4907 - 01 h m . s + ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . α Per14802 97 11072 kap For 02 h m . s - ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ - . . h m . s - ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . . h m . s + ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ c - . d - . d - - 50–150 W03 -82558 355 46816 LQ Hya 09 h m . s - ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . . h m . s - ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ c - . d - . d - - 50–150 W03 -111395 486.1 62523 - 12 h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h e G e m i n i D ee p P l a n e t S u r v e y5 TABLE 1 —
Continued
Names α δ
Spectral H Dist. a µ α cos δ a µ δ a [Fe/H] [Fe/H] Age Age NotesHD GJ HIP Other (J2000) (J2000) Type (mag) (pc) (mas/yr) (mas/yr) Ref. (Myr) Ref.- 507.1 65016 - 13 h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . . h m . s - ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ . . h m . s + ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . . h m . s - ◦ ′ . ′′ - . - . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ - . . h m . s - ◦ ′ . ′′ - . - . h m . s + ◦ ′ . ′′ e . e - . e -0.70 N04 50–150 W03 m167605 - 89005 LP Dra 18 h m . s + ◦ ′ . ′′ - . . h m . s + ◦ ′ . ′′ . . h m . s + ◦ ′ . ′′ . . h m . s - ◦ ′ . ′′ . - . β Pic201651 - 104225 - 21 h m . s + ◦ ′ . ′′ . . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ b . b . b - - 50–150 W03,M01a,C05 LA213845 863.2 111449 ups Aqr 22 h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . . h m . s - ◦ ′ . ′′ . - . h m . s + ◦ ′ . ′′ . . h m . s - ◦ ′ . ′′ . . R EFERENCES . — (Z) B. Zuckerman, private communication.; (B99) Barrado y Navascués et al. 1999; (C05) Carpenter et al. 2005; (C01) Cayrel de Strobel et al. 2001; (C97) Cayrel de Strobel et al. 1997; (C04) Clem et al. 2004; (F95) Favata et al. 1995; (F98) Favata et al. 1998; (F04) Fuhrmann 2004; (G00) Gaidos et al. 2000; (G98)Gaidos 1998; (G01) Gerbaldi et al. 2001; (G03) Gray et al. 2003; (H98) Huensch et al. 1998; (H99) Hünsch et al. 1999; (K01) Kirkpatrick et al. 2001; (K02) Kotoneva et al. 2002; (L99) Lachaume et al. 1999; (L06) López-Santiago et al. 2006; (L05) Lowrance et al. 2005; (M06) Makarov 2006; (M03) Mohanty & Basri 2003; (M01a) Monteset al. 2001b; (M01b) Montes et al. 2001a; (N04) Nordström et al. 2004; (R05) Rieke et al. 2005; (S05) Saffe et al. 2005; (S04a) Santos et al. 2004; (S93) Soderblom & Mayor 1993; (S00) Song et al. 2000; (S04b) Song et al. 2004; (S06) Sousa et al. 2006; (T05a) Takeda & Kawanomoto 2005; (T05b) Taylor 2005; (V04) Valdes et al. 2004;(V05) Valenti & Fischer 2005; (W03) Wichmann et al. 2003; (W04) Wright et al. 2004; (W05) Wyatt et al. 2005; (Z01) Zuckerman et al. 2001; (Z04b) Zuckerman & Song 2004b; (Z04a) Zuckerman & Song 2004a; (Z04c) Zuckerman et al. 2004; (Z06) Zuckerman et al. 2006N
OTE . — Star is a member of ( α Per) α Persei; (AB Dor) AB Doradus; ( β Pic) β Pictoris; (Ca-Near) Carina-Near; (Her-Lyr) Hercules-Lyra; (LA) Local association; (LA (B4)) Local association, subgroup B4; (U Ma) Ursa Major. If a question mark precedes the association, the membership is doubtful or based on kinematics only. An“m” indicates that the star is a multiple, see § 4.3 for more detail. a From the Hipparcos catalog (Perryman & ESA 1997), unless stated otherwise. b From Montes et al. (2001b). c From the Tycho catalog (Høg et al. 1997). d From the Tycho-2 catalog (Høg et al. 2000). e From Fekel et al. (2005).
Lafrenière et al.From this preliminary compilation, we have retained onlystars with a distance smaller than 35 pc, and we have excludedstars of declination below - ◦ since observations were to bemade from the Gemini North observatory. Finally, we havefurther excluded stars indicated to be multiple in Zuckerman& Song (2004b). This procedure yielded a list of slightly over100 target stars, of which 85 were actually observed. Theproperties of these 85 stars are presented in Table 1 and Fig-ure 1. The median spectral type of our sample is K0, the me-dian H magnitude is 5.75, the median distance is 22 pc, themedian proper motion amplitude is 240 mas yr - , and the me-dian [Fe/H] is 0.00 dex (standard deviation of 0.21 dex).Despite our effort to select only single stars, our observa-tions show that 16 of the 85 target stars are close double ortriple systems; this is indicated in the last column of Table 1.A thorough review of the literature revealed that 11 of thesewere known at the time the target list was compiled, two ofwhich are astrometric multiples that had never been resolvedprior to our observations (HD 14802 and HD 166181). Fiveother multiple systems were resolved with AO only after thetarget list was compiled (HD 77407, HD 129333, HD 135363,HD 160934, and HD 220140). Finally, the star HD 213845 isreported to be part of a binary system for the first time here.The multiple systems observed are discussed further in §4.3.Age estimates for the stars in our sample, needed to convertthe observed contrasts into mass detection limits using evo-lution models of giant planets, are reported in Table 1 alongwith the references used for their determination. Wheneverpossible, we have used ages stated explicitly in the literatureor the age of the association to which a star belongs. Whenno specific age estimate was available for stars taken fromWichmann et al. (2003), ages of 10–50 Myr or 50–150 Myrwere assigned to the stars having a lithium abundance aboveor comparable to that of the Pleiades, respectively. For otherstars that have lithium and/or X-ray measurements, ages wereestimated from a comparison of the Li I 6708 Å equivalentwidth and/or the ratio of the X-ray to bolometric luminositywith Figures 3 and/or 4 of Zuckerman & Song (2004b) respec-tively. When lithium or X-ray measurements were not avail-able, the kinematic ages were used as lower limits while theages derived from the chromospheric activity index, log R HK ,were used as upper limits, as Song et al. (2004) showed thatthe latter ages tend to be systematically higher than those de-rived from lithium abundance or X-ray emission. When onlythe value of log R HK was available, the calibration of Donahue(1993) was used to obtain an age estimate. Finally, whenonly kinematics measurements were available for a given star,an age of 100–5000 Myr or 50–5000 Myr was assigned ifthe star is a possible member of the Local Association or theIC 2391 supercluster respectively. OBSERVATIONS AND IMAGE PROCESSING3.1.
Data acquisition and observing strategy
All observations were obtained at the Gemini North tele-scope with the Altair adaptive optics system (Herriot et al.2000) and the NIRI camera (Hodapp et al. 2003) (pro-grams GN-2004B-Q-14, GN-2005A-Q-16, GN-2005B-Q-4,GN-2006A-Q-5, and GN-2006B-Q-5). The f /
32 camera wasused, yielding 0.022 ′′ pixel - and a field of view of 22 ′′ × ′′ .The field lens of Altair, which improves the off-axis adap- It is assumed that any planet and its primary star would be coeval. This calibration is given explicitly in Henry et al. (1996). tive optics correction, was not used for any observation as itwas not available for the first epoch observations. Becauseit introduces an undetermined field distortion, having usedthe field lens for the second epoch observations only wouldhave complicated or prevented verification of the physical as-sociation of companion candidates identified in the first epochobservations. The observations were obtained in the narrowband filter CH4-short (1.54–1.65 µ m), for the following rea-son. According to evolution models (e.g. Baraffe et al. 2003),planetary mass objects older than 10-20 Myr should have aneffective temperature below 1000 K. Because of the largeamounts of methane and the increased collision induced ab-sorption by H in their atmosphere, the near-infrared K -bandflux of such objects is largely suppressed. It is thus moreefficient to search for giant planets in either the J or the H band; the latter was preferred in this study because higherStrehl ratios are achieved at longer wavelengths. As the bulkof the H -band flux of cool giant planets is emitted in a nar-row band centered at ∼ µ m because of important absorp-tion by methane beyond 1.6 µ m, it is even more efficient tosearch for these planets using the CH4-short filter, which iswell matched to the peak of the emission. Based on evolutionmodels and synthetic spectra of giant planets (Baraffe et al.2003), it is expected that the mean flux density of a planet inthe NIRI CH4-short filter be between 1.5 and 2.5 times higher(0.44–1.0 mag brighter) than in the broad band H filter, de-pending on the specific age and mass of the planet. Thesefactors are consistent with the factors 1.6-2.0 (0.5–0.75 mag)calculated from the observed spectra of T7–T8 brown dwarfs,which have T eff ∼
800 K.The angular differential imaging (ADI, Marois et al. 2006)technique was used to suppress the PSF speckle noise and im-prove our sensitivity to faint companions. This technique con-sists of acquiring a sequence of many exposures of the targetusing an altitude/azimuth telescope with the instrument rota-tor turned off (at the Cassegrain focus) to keep the instrumentand telescope optics aligned. This is a very stable configu-ration and ensures a high correlation of the sequence of PSFimages. This setup also causes a rotation of the field of view(FOV) during the sequence. For each target image in such asequence, it is possible to build a reference image from othertarget images in which any companion would be sufficientlydisplaced due to FOV rotation. After subtraction of the ref-erence image, the residual images are rotated to align theirFOV and co-added. Because of the rotation, the residual PSFspeckle noise is averaged incoherently, ensuring an ever im-proving detection limit with increasing exposure time. It hasbeen shown that, for ADI with Altair/NIRI, the subtraction ofan optimized reference PSF image from a target image cansuppress the PSF speckle noise by a factor of ∼
12, and thata noise suppression factor of ∼
100 is achieved for the com-bination of 90 such difference images (Lafrenière et al. 2007;Marois et al. 2006).An individual exposure time of 30 seconds was chosenfor all targets. This exposure time is long enough so that, atlarge separation, faint companion detection is limited by skybackground noise rather than read noise, and short enoughso that the radius of the region affected by saturation andnon-linearity of the detector typically does not exceed 0.5 ′′ .The nominal observing sequence consisted of 90 images,but oftentimes a few images had to be discarded due to briefperiods of very bad seeing, loss of tracking, or the advent ofclouds. No dithering was made during the main observingsequence to ensure a high correlation of the PSF images; flat-he Gemini Deep Planet Survey 7 TABLE 2GDPS
OBSERVATION LOG
Name Date Number of Strehl FOV rotation Saturationexposures (%) (deg) radius ( ′′ ) a HD 166 2005/08/25 83 5-8 55 0.982006/07/18 83 7-10 81 0.78HD 691 2005/08/10 90 13-17 70 0.432006/09/18 117 16-30 88 0.44HD 1405 2004/08/22 90 4-10 17 0.532005/08/04 90 6-18 69 0.40HD 5996 2005/08/12 90 18-20 24 0.502006/09/25 90 15-17 21 0.50HD 9540 2005/08/14 90 16-19 25 0.552006/09/28 45 14-17 11 0.61HD 10008 2005/08/10 90 18-20 36 0.51GJ 82 2005/08/31 90 10-12 27 0.28HD 14802 2005/08/20 90 - 23 1.09HD 16765 2005/09/10 90 14-17 45 0.72HD 17190 2005/08/24 90 13-30 108 0.52HD 17382 2004/12/22 66 15 68 0.552005/09/11 90 19-23 104 0.52HD 17925 2004/11/04 83 & b
29 0.66HD 18803 2004/12/24 90 7-14 99 0.702005/09/12 78 17-18 108 0.64HD 19994 2005/08/31 90 - 44 0.832006/10/01 57 - 27 0.72HD 20367 2005/10/02 90 12-14 67 0.702E 759 2005/10/17 59 7-10 31 0.22HD 22049 2005/09/08 90 - 32 2.06HIP 17695 2005/09/13 89 20-20 45 0.24HD 25457 2005/10/02 90 - 43 0.96HD 283750 2004/10/24 90 15 99 0.542005/10/04 87 19-23 101 0.59HD 30652 2005/09/12 52 - 35 1.86GJ 182 2004/11/05 90 16-20 31 0.372005/10/17 33 11-11 29 0.39GJ 234A 2005/11/05 72 16 34 0.42GJ 281 2005/03/25 67 9-10 49 0.522006/02/12 25 8-9 11 0.47GJ 285 2005/03/18 20 - 10 0.452006/02/12 90 4-5 73 0.55HD 72905 2005/04/23 84 7 25 0.87HD 75332 2005/04/24 89 & b
27 0.592006/12/20 16 & b
11 0.59HD 77407 2005/04/26 84 16-19 33 0.61HD 78141 2004/12/21 85 14-16 19 0.55HD 82558 2005/04/18 90 - 30 0.61HD 82443 2004/12/25 75 18 28 0.61GJ 393 2005/04/20 90 13-15 44 0.55HD 90905 2005/03/18 90 13-18 47 0.612006/04/11 35 13-15 14 0.55HD 91901 2005/04/29 71 9 22 0.44HD 92945 2005/05/26 85 15-16 19 0.612006/05/16 10 10-11 2 0.53HD 93528 2005/04/30 86 - 26 0.39GJ 402 2005/04/26 79 12-16 37 0.392006/02/16 60 6-10 33 0.35HD 96064 2005/04/19 89 21-23 37 0.502006/03/05 90 13-19 36 0.50HD 97334 2005/04/18 90 16-17 54 0.70HD 102195 2005/04/24 91 20-21 54 0.412006/03/18 82 12-18 30 0.39HD 102392 2005/04/23 89 19-24 32 0.392006/03/12 90 9-13 31 0.40HD 105631 2005/05/29 90 14-19 45 0.55HD 107146 2005/05/30 90 21-26 71 0.57HD 108767B 2005/04/22 90 14 27 0.432006/02/16 43 10-11 14 0.41HD 109085 2005/05/26 90 - 22 1.092006/03/12 15 - 3 1.09BD+60 1417 2005/04/18 90 18-23 24 0.262006/04/11 63 12 19 0.24HD 111395 2005/04/19 89 & b
120 0.77HD 113449 2005/06/01 47 10-20 37 0.52GJ 507.1 2005/06/07 87 5-7 61 0.44HD 116956 2005/05/29 90 5-14 27 0.55
Lafrenière et al.
TABLE 2 —
Continued
Name Date Number of Strehl FOV rotation Saturationexposures (%) (deg) radius ( ′′ ) a & b
122 0.83HD 135363 2005/04/18 87 14-15 19 0.482006/02/16 60 8-9 14 0.44HD 139813 2005/05/30 90 & b
20 0.57HD 141272 2005/04/19 90 18-19 47 0.552006/03/12 42 13 20 0.56HD 147379B 2005/04/18 90 17-17 22 0.50GJ 628 2005/04/17 90 11 29 0.702006/04/11 40 9-14 13 0.66HIP 81084 2005/04/19 73 17-18 30 0.332006/05/15 90 8-13 31 0.22HD 160934 2005/04/18 84 17-24 24 0.352006/09/17 14 12-14 4 0.34HD 162283 2005/04/20 120 15-19 45 0.382006/09/16 100 27-29 31 0.33HD 166181 2005/04/17 90 16 76 0.592006/09/18 45 18-21 37 0.48HD 167605 2005/05/27 90 20 22 0.39HD 187748 2005/05/25 97 15-19 30 0.662006/09/15 75 & b
21 0.50GJ 791.3 2005/05/26 87 9-19 54 0.42HD 197481 2005/07/29 68 6-10 21 0.87HD 201651 2005/06/27 90 18-23 21 0.382006/09/14 30 19-21 7 0.38HD 202575 2005/07/16 90 17-23 75 0.572006/09/14 30 16-18 9 0.56GJ 4199 2004/08/23 65 10-13 118 0.442005/08/04 90 15-23 136 0.39HD 206860 2005/08/10 34 & b
56 0.772006/06/26 60 & b
80 0.61HD 208313 2005/06/27 90 23-23 67 0.462006/06/25 89 14-22 66 0.55V383 Lac 2005/07/26 66 13-17 28 0.422006/06/30 77 15-18 27 0.32HD 213845 2005/08/24 90 - 26 0.812006/07/06 90 - 24 0.83GJ 875.1 2005/08/10 90 16-18 69 0.332006/07/07 79 7-17 61 0.31GJ 876 2005/08/21 82 9-16 28 0.68GJ 9809 2005/08/04 90 18-20 25 0.312006/09/14 120 25-27 31 0.22HD 220140 2005/08/05 90 16-18 21 0.592006/07/16 82 7-9 19 0.63HD 221503 2005/08/31 90 21-22 28 0.52GJ 900 2004/08/24 90 15-21 17 0.462005/09/08 90 16-22 46 0.42GJ 907.1 2005/09/07 65 5-15 22 0.372006/07/17 44 8 16 0.31 a Radius at which the PSF radial intensity profile reaches 75% of the detector well capacity. b Only a lower estimate of the Strehl ratio can be obtained as the PSF peak is in the non-linear regime or sligthly saturated. field errors, bad pixels, and cosmic ray hits are naturally av-eraged/removed with ADI because of the FOV rotation. ThePSF centroid was found to wander over the detector by typ-ically 2-5 pixels throughout an observing sequence becauseof mechanical flexure and differential refraction between thewavefront sensing and science wavelengths; for a handful oftargets the variation slightly exceeded 10 pixels. Short unsat-urated exposures were acquired before and after the main se-quence of (saturated) images for photometric calibration andStrehl ratio estimation; these observations were acquired in sub-array mode (256 ×
256 or 512 ×
512 pixels), for which theminimum exposure time is shorter. Typically, an unsaturatedsequence consisted of five exposures each obtained at a differ-ent dither position. The unsaturated observations are missingfor a few targets as they were either skipped in the executionof the program, or they turned out to be saturated despite us-ing the shortest possible exposure time. Table 2 summarizesall observations. The last column of the table (“saturation ra-dius”) indicates the separation at which the radial profile ofthe PSF reaches 75% of the detector full well capacity; linear-he Gemini Deep Planet Survey 9 F IG . 2.— Illustration of the ADI noise attenuation process. Panel (a) shows an original 30-s image of the young star HD 691 after subtraction of an azimuthallysymmetric median intensity profile, panels (b) and (c) both show, with a different intensity scale, the corresponding residual image after ADI subtraction usingthe LOCI algorithm, and panel (d) shows the median combination of 117 such residual images. Display intensity ranges are ± × - and ± - of stellarPSF peak for the top and bottom rows respectively. Each panel is 10 ′′ on a side. The diffraction spikes from the secondary mirror support vanes and the centralsaturated region are masked. The faint point source ( ∆ m = 14 .
9) visible in panel (d) at a separation of 2.43 ′′ and P.A. of 7.3 ◦ could not have been detectedwithout ADI processing. ity should be better than 1% at this level (Hodapp et al. 2003).We have not analyzed the data inside this separation; pointsources located at least one PSF full-width-at-half-maximum(FWHM) past this separation can be detected in our analysis,provided that their brigthness is above the detection limit.3.2. Data reduction
For each sequence of short unsaturated exposures, a skyframe was constructed by taking the median of the imagesobtained at different dither positions; this sky frame was sub-tracted from each image. The images were then divided bya flat field image. The PSFs of a given unsaturated sequencewere registered to a common center and the median of the im-age sequence was obtained. The center of the PSFs were de-termined by fitting a 2-dimensional Gaussian function. As anindication of the quality of an observing sequence, the Strehlratio was calculated by comparing the peak pixel value of theobserved PSF image with that of an appropriate theoreticalPSF. The calculated Strehl ratio values are reported in Table 2; two values are indicated for a target when unsaturated datawere obtained before and after the main saturated sequence.Strehl ratios were typically in the range 10–20%.Images of the main saturated sequence were first dividedby a flat field image. Bad and hot pixels, as determinedfrom analysis of the flat field image and dark frame respec-tively, were replaced by the median value of neighboring pix-els. Field distortion was corrected using an IDL procedureprovided by the Gemini staff (C. Trujillo, private communi-cation) and modified to use the IDL interpolate function withcubic interpolation. The plate scale and field of view orien-tation for each image were obtained from the FITS headerkeywords.For each sequence of saturated images, the stellar PSF ofthe first image was registered to the image center by maxi-mizing the cross-correlation of the PSF diffraction spikes withthemselves in a 180-degree rotation of the image about itscenter. The stellar PSF of the subsequent images was regis-tered to the image center by maximizing the cross-correlation0 Lafrenière et al.of the PSF diffraction spikes with those in the first image.Prior to shifting, the 1024 × × Photometric calibration and uncertainty
As the stellar PSF peak is saturated for the main sequence ofimages, and since much image processing is done to subtractthe stellar PSF from each image, special care must be takento calibrate the photometry of the residual images and ensurethat the contrast limits calculated are accurate.When the PSF peak is saturated, relative photometry canbe calibrated by scaling the stellar flux measured in the un-saturated images obtained before and/or after the saturatedsequence according to the ratio of the exposure times of thesaturated and unsaturated images. However, the accuracy ofthis calibration method is affected by the (unknown) varia-tions in Strehl ratio, hence of the peak PSF flux, that may haveoccurred between the saturated and unsaturated observations.To mitigate this problem, the calibration approach we adoptedrelies on a sharp ghost artifact located ( + . ′′ , - . ′′ ) fromthe PSF center in the ALTAIR/NIRI images. Since the inten-sity of this ghost artifact is proportional to the PSF intensity, itcan be used to infer the peak flux of a saturated PSF. This wasverified for all sequences for which both unsaturated and satu-rated data were available. First, the stellar flux was measuredin the unsaturated images using a circular aperture of diam-eter equal to the FWHM of the PSF. When unsaturated datawere acquired both before and after the saturated sequence,the mean of the two values was used. Then the flux of theghost artifact in the same aperture was measured for each im-age of the saturated sequence. The median of these values,scaled according to the ratio of the exposure times of the satu-rated and unsaturated images, was then compared to the stel-lar flux, and the process was repeated for all sequences thatinclude both saturated and unsaturated data. Similar valueswere found for all sequences; the mean ratio of the flux of theghost over that of the PSF peak was found to be 6 . × - ,with a standard deviation of 0 . × - . Comparisons of theflux of background stars bright enough to be visible in eachindividual image of a sequence with the flux of the ghost inthe corresponding images also confirmed that the intensity ofthe ghost is indeed directly proportional to the intensity ofoff-axis sources.The procedure used for calibrating the photometry was thefollowing. The flux of the ghost was measured for each im- C o rr e c t i on f a c t o r F IG . 3.— Typical values of f sub ( solid line ), f aniso ( dotted line ), and f sm ( dashed line ) as a function of angular separation. The curves shown are forthe target HD 166181. age of a sequence and the median of these values, divided bythe ratio quoted above, was taken to represent the peak stellarPSF flux, F ⋆ . This calibration method should be more accu-rate than the one based solely on unsaturated data obtainedbefore and/or after the saturated sequence because the me-dian ghost flux is affected in the same way as the median ofall residual images by the variations of Strehl ratio that mayhave occurred during the sequence of saturated images or be-tween the saturated and unsaturated measurements. For thisreason, this calibration was used even for the sequences forwhich unsaturated data were available.Observations obtained with ALTAIR without the field lenssuffer from important off-axis Strehl degradation because ofanisoplanatism; this degradation must be taken into accountwhen calculating contrast. Unfortunately, it is virtually im-possible to quantify the specific degradation pertaining to ourdata as there are no bright reference off-axis point sourcesavailable for every sequence of images. Instead, we haveused the average anisoplanetism Strehl ratio degradation for-mula indicated on the ALTAIR webpage , which is f aniso ( θ ) ≡ S ( θ ) / S = e - ( θ / . , where S ( θ ) is the Strehl ratio at angularseparation θ , expressed in arcseconds, and S is the on-axisStrehl ratio. This factor was used to correct the noise and theflux of faint point sources measured in the residual images.As explained in Lafrenière et al. (2007), while the subtrac-tion of an optimized reference PSF obtained using the LOCIalgorithm leads to better signal-to-noise (S/N) ratios, it re-moves partially the flux of the point sources sought after. Thisflux loss must be accounted for when calculating contrast.This is done by calculating the normalized residual intensity, f sub , of artificially implanted point sources after execution ofthe subtraction algorithm; the method used is described in§4.3 of Lafrenière et al. (2007). Then using flux measure-ments made in the residual image, the factor f sub is used toinfer the true flux of a point source, i.e. that before executionof the subtraction algorithm.Another effect that must be taken into account for ADIdata is the azimuthal smearing of an off-axis point sourcethat occurs as the field of view rotates during an integration;this causes a fraction of the source’s flux to fall outside ofthe circular aperture used for photometric measurements.The amount of flux loss in the aperture was calculated for each he Gemini Deep Planet Survey 11 TABLE 3P
HOTOMETRIC UNCERTAINTIES
Sep. ( ′′ ) < - -
10 10 - > σ (mag) 0.07 0.12 0.15 0.26 0.39 sequence of images as follows. For a given angular separa-tion and for each image of a sequence, a copy of the unsatu-rated PSF was smeared according to its displacement duringan integration. When unsaturated data were unavailable, a 2DGaussian of the appropriate FWHM was used in place of theunsaturated PSF. The median of these smeared PSFs was ob-tained and the flux in a circular aperture was measured. Thisflux was divided by the flux of the original PSF in the sameaperture to obtain the smearing factor f sm , which is used tocorrect the flux or noise measured in the images.Given all of these considerations, the contrast at angularseparation θ was calculated as C ( θ ) = F ( θ ) f aniso ( θ ) f sm ( θ ) f sub ( θ ) × F ⋆ , (1)where F ( θ ) is either the noise or the flux of a point sourcein a circular aperture of diameter equal to one PSF FWHM,at angular separation θ , in the residual image. Note that thecontrast in the equation above is defined such that a faintercompanion, or a smaller residual noise, has a smaller con-trast value. Eq. (1) was used for all contrast calculations inthe present work. Typical correction factors as a function ofangular separation are shown in Figure 3.An estimate of the photometric accuracy resulting from theentire process was obtained by calculating the mean absolutedifference between the magnitudes calculated at two epochsfor every faint background star that was observed twice (see§4.2); this mean absolute difference was taken to represent √ ′′ . For com-pleteness, it is noted that a higher photometric uncertainty, byabout 0.08 mag, results when the unsaturated data obtainedbefore and/or after the main sequence of saturated images areused to determine F ⋆ , rather than the median flux of the ghostartifact, justifying our choice to use the calibration based onthe flux of the ghost for all sequences. RESULTS4.1.
Detection limits
Detection limits are based on a measure of the noise in theresidual images. To calculate this noise, the residual imageswere first convolved by a circular aperture of diameter equalto one PSF FWHM, which is typically ∼ ′′ , and the noiseas a function of angular separation from the image center, F ( θ ), was determined as the standard deviation of the pixelvalues in an annulus of width equal to one PSF FWHM. Asshown in Lafrenière et al. (2007) and Marois et al. (2007b),the noise in an ADI residual image has a distribution similarto a Gaussian; using a 5 σ detection threshold is thus appropri-ate for our data to limit the number of false positives. Giventhat a residual image typically contains ∼ × resolution elements, roughly 0.1 false positive per target is expected onaverage. Because of the underlying noise in a residual im-age, some sources near the detection threshold might not bedetected. From Gaussian statistics, the probability that theresidual signal underlying a source is below 0 σ , - σ , or - σ is 50%, 16%, or 2.3%, respectively. Our detection complete-ness for sources whose true intensities are 5 σ , 6 σ , or > σ isthus 50%, 84%, or > σ threshold could be detected as well. These effects will betaken into account appropriately in the statistical analysis ofthe results presented in §5.The detection limits achieved for all target stars, expressedin magnitude difference, are presented in Table 4. The lasttwo lines of this table present the median and best contrast,over the 85 observations, achieved at each angular separation.The median detection limits in magnitude difference are 9.5at 0.5 ′′ , 12.9 at 1 ′′ , 15.0 at 2 ′′ , and 16.5 at 5 ′′ . The detec-tion limits are presented graphically in Figure 4 for the starsHD 208313, HD 166181, and GJ 507.1, which are representa-tive of poor, median, and good contrast performance, respec-tively.For consistency we have verified the validity of these detec-tion limits by implanting fiducial sources in the sequence oforiginal images and then processing the data as described in§3.2. An example, incorporating artificial sources at the 5 σ and 10 σ levels at various separations, is shown in Figure 5 forthe stars HD 208313, HD 166181, and GJ 507.1. As visible inthis figure, sources exactly at our detection limits can indeedbe detected with the expected completeness level.One must resort to evolution models of giant planets toconvert the detection limits mentioned above into masslimits. Traditionally, such evolution models have assumedarbitrary initial conditions for the planets (e.g. Baraffe et al.2003; Burrows et al. 1997), with the caution that their resultsdepend on the specific initial conditions adopted for agesbelow a few million years (Baraffe et al. 2002). Recentevolution models (Marley et al. 2007) that incorporateinitial conditions calculated explicitly for planets formedthrough core accretion indicate that it may in fact take asmuch as 10–100 Myr before the planets “forget” their initialconditions; the effect being more important for more massiveplanets. Nevertheless, given the typical ages of our targetstars (50–300 Myr) and the good contrast limits we havereached, the different evolution models should yield similarmass detection limit estimates. As a simple example, considera contrast of 12.9 mag in the NIRI CH4-short filter around aK0 star (typical at a separation of 1 ′′ ). The “hot start” modelsof Baraffe et al. (2003) would give masses of 2.6 M Jup and3.9 M Jup at 50 Myr and 100 Myr, respectively, while the “coreaccretion” models of Marley et al. (2007) would give massesof ∼ M Jup and ∼ M Jup , respectively . The differencebetween the models would be smaller for smaller masses(better contrast limits (i.e. beyond ∼ ′′ ) and/or greater ages),while it would be larger for larger masses (worse contrastlimits and/or smaller ages). In this work, keeping the lattercaveat in mind, we have used the COND evolution models ofBaraffe et al. (2003), for which absolute H -band magnitudesas a function of mass and age are readily available. Thefollowing procedure was used to estimate the contrast, in the For this simple calculation, it was assumed that the luminosity ratiosbetween the “hot start” and “core accretion” models were representative of the H -band magnitude differences. C on t r a s t li m i t ( D m ag , s )
25. 50. 75. 100. 125. 150. 175. 200.Projected physical separation (AU) at 22 pc 1.01.52.03.04.05.07.010.012.0 M a ss li m i t f o r a M y r K p r i m a r y ( M J up ) F IG . 4.— Survey detection limits in difference of magnitude (in the NIRI CH4-short filter) between an off-axis point source and the target star, at the 6 σ level.The top, middle, and bottom curves are respectively for the targets GJ 507.1, HD 166181, and HD 208313, which are representative of poor, median, and goodperformance reached by the survey. Companion candidates identified around targets of galactic latitude | b | <
15 are shown by + symbols, while those identifiedaround targets with | b | ≥
15 are shown by × symbols. The two filled circles near (2.6,8.6) indicate the components of the binary brown dwarf companion toHD 130948. The fiducial point sources shown in Fig. 5 are marked with triangles. The top and right axes show, for reference only, the projected separation in AUand the detection limits in M Jup that would apply for a 100 Myr old K0 star located 22 pc away.TABLE 4GDPS
DETECTION LIMITS a Name 0.50 ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ HD 166 - - - 12.5 13.1 13.9 14.9 15.4 15.9 16.5 16.9 17.3 17.3HD 691 - 11.1 12.1 13.2 14.1 14.7 15.6 15.9 16.2 16.6 16.7 16.6 16.3HD 1405 9.2 10.5 11.4 12.7 13.5 14.0 14.8 15.3 15.7 16.0 16.1 16.1 15.8HD 5996 - 10.8 12.0 13.2 14.1 14.6 15.4 15.8 16.1 16.5 16.6 16.6 16.4HD 9540 - - 11.8 13.1 14.0 14.5 15.4 16.0 16.4 17.0 17.3 17.6 17.5HD 10008 - 10.0 11.2 12.4 13.2 13.8 14.7 15.2 15.6 16.2 16.5 16.5 16.3GJ 82 8.9 9.5 10.5 11.8 12.5 13.2 13.7 14.3 14.6 14.9 15.0 14.8 14.6HD 14802 - - - - 11.8 12.4 13.3 14.0 14.7 15.8 16.8 17.4 17.9HD 16765 - - - 13.0 13.9 14.5 15.3 15.8 16.2 16.9 17.4 17.5 17.6HD 17190 - 10.5 12.2 13.7 14.2 14.8 15.5 15.9 16.3 16.6 16.8 16.6 16.2HD 17382 - 10.8 12.0 13.3 14.1 14.6 15.4 15.9 16.3 16.8 17.0 17.0 16.7HD 17925 - - 11.9 13.6 14.6 15.4 16.2 16.8 17.1 17.6 17.7 17.7 17.4HD 18803 - - 11.3 12.9 13.8 14.5 15.5 16.0 16.5 16.8 17.1 17.2 16.9HD 19994 - - - 13.5 14.3 15.0 15.8 16.4 16.7 17.4 17.8 18.3 18.4HD 20367 - - - 11.6 12.2 12.8 13.9 14.4 14.8 15.6 16.0 16.3 16.12E 759 8.6 9.4 9.9 11.0 11.8 12.2 13.0 13.4 13.6 13.9 14.0 13.9 13.6HD 22049 - - - - - - - 15.9 16.5 17.3 17.7 18.5 18.9HIP 17695 10.0 10.8 11.8 12.8 13.6 14.2 14.8 15.1 15.3 15.7 15.7 15.6 15.3HD 25457 - - - - 12.5 13.1 13.9 14.8 15.2 16.0 16.6 17.0 17.0HD 283750 - - 12.2 13.4 14.2 15.1 15.9 16.4 16.8 17.2 17.2 17.1 16.7HD 30652 - - - - - - 14.9 15.5 15.9 16.7 17.3 18.2 18.6GJ 182 10.0 10.5 11.9 13.1 14.0 14.7 15.4 15.8 16.1 16.4 16.5 16.4 16.2GJ 234A 9.5 10.1 11.2 12.3 13.3 13.9 14.6 15.1 15.4 15.9 16.2 16.3 16.1GJ 281 - 9.0 10.4 12.0 12.9 13.5 14.3 14.6 15.0 15.3 15.3 15.4 15.2GJ 285 - 8.0 10.1 11.6 12.6 13.3 13.8 14.5 14.9 15.5 15.8 15.9 15.8HD 72905 - - - 11.2 12.5 13.1 14.2 14.9 15.4 16.3 16.7 17.4 17.7HD 75332 - - 10.8 12.3 13.0 13.9 14.9 15.5 15.7 16.6 17.1 17.4 17.3HD 77407 - - 10.3 11.4 12.3 13.0 14.0 14.8 15.0 15.7 16.0 16.3 16.2HD 78141 - - 11.5 13.0 13.7 14.5 15.4 15.8 16.1 16.5 16.6 16.5 16.3HD 82558 - - 11.5 12.9 13.8 14.4 15.4 15.9 16.1 16.6 16.8 17.0 16.7HD 82443 - - 11.5 13.0 14.1 14.8 15.9 16.4 16.8 17.2 17.5 17.7 17.5 he Gemini Deep Planet Survey 13
TABLE 4 —
Continued
Name 0.50 ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ ′′ GJ 393 - - 11.8 13.3 14.1 14.6 15.6 16.0 16.2 16.7 16.8 16.9 16.8HD 90905 - - 11.4 12.7 13.7 14.1 15.1 15.7 16.1 16.5 16.6 16.6 16.4HD 91901 - 9.2 10.0 11.4 12.1 12.8 13.6 14.1 14.4 14.9 14.8 14.8 14.6HD 92945 - - 10.8 12.1 13.0 13.8 14.6 15.1 15.5 15.9 16.1 16.3 16.1HD 93528 8.5 9.3 10.2 11.6 12.6 13.3 14.2 14.8 15.0 15.5 15.7 15.9 15.7GJ 402 8.4 9.2 10.5 11.6 12.5 13.1 14.0 14.5 14.9 15.4 15.4 15.6 15.3HD 96064 - 10.9 12.3 13.5 14.3 14.9 15.6 16.1 16.3 16.6 16.8 16.8 16.6HD 97334 - - - 13.6 14.7 15.1 16.0 16.4 16.7 17.2 17.4 17.5 17.3HD 102195 9.8 11.2 12.2 13.3 14.1 14.7 15.4 15.9 16.1 16.5 16.6 16.6 16.3HD 102392 9.5 10.3 11.4 12.6 13.5 13.9 14.7 15.3 15.6 16.1 16.3 16.3 16.2HD 105631 - - 11.7 12.8 13.5 14.2 15.1 15.5 16.0 16.4 16.7 16.8 16.5HD 107146 - - 11.7 12.5 13.5 14.0 15.0 15.4 15.8 16.2 16.5 16.5 16.3HD 108767B 8.4 9.7 10.6 11.9 12.8 13.5 14.3 14.9 15.1 15.7 15.8 16.0 15.7HD 109085 - - - - 13.4 14.0 14.9 15.8 16.3 17.2 17.7 18.3 18.5BD+60 1417 10.0 11.1 12.0 13.0 13.8 14.2 14.7 15.0 15.3 15.5 15.5 15.4 15.1HD 111395 - - - 13.4 14.3 15.0 15.9 16.4 16.7 17.2 17.4 17.6 17.3HD 113449 - - 11.5 12.6 13.7 13.9 14.9 15.4 15.7 16.3 16.5 16.6 16.4GJ 507.1 - 9.6 10.4 11.5 12.2 12.9 13.7 14.3 14.6 15.0 15.2 15.2 14.9HD 116956 - - 11.3 12.7 13.5 14.2 15.1 15.7 16.0 16.5 16.7 16.8 16.6HD 118100 8.4 9.4 10.5 11.6 12.3 12.8 13.5 14.0 14.1 14.4 14.5 14.4 14.2GJ 524.1 10.1 11.0 12.0 13.0 13.6 14.2 14.9 15.2 15.4 15.4 15.5 15.4 15.0HD 124106 - 10.3 11.6 13.0 13.8 14.4 15.4 15.7 15.9 16.4 16.7 16.8 16.6HD 125161B 10.5 11.3 12.4 13.6 14.3 14.6 15.4 15.8 16.0 16.2 16.4 16.3 16.1HD 129333 - 10.7 11.7 13.2 13.9 14.4 15.3 15.7 16.2 16.4 16.7 16.7 16.5HD 130004 - - 12.0 13.1 14.1 14.5 15.3 15.8 16.1 16.5 16.7 16.7 16.4HD 130322 - 11.1 12.1 13.2 13.9 14.3 15.2 15.6 15.9 16.3 16.4 16.4 16.2HD 130948 - - - 12.4 13.2 13.8 14.7 15.4 15.7 16.5 16.9 17.3 17.3HD 135363 - 9.2 10.9 12.3 13.1 13.7 14.6 15.1 15.3 15.5 15.7 15.6 15.4HD 139813 - - 10.3 11.2 11.9 12.6 13.6 14.3 14.9 15.7 16.1 16.3 16.1HD 141272 - - 12.2 13.7 14.4 15.0 15.8 16.2 16.5 16.9 16.9 17.1 16.9HD 147379B - 10.0 11.3 12.8 13.5 14.1 15.0 15.3 15.6 15.8 16.0 16.0 15.7GJ 628 - - 10.4 12.2 13.0 13.7 14.6 15.2 15.7 16.2 16.6 16.9 16.7HIP 81084 9.5 10.3 11.4 12.3 13.0 13.5 14.0 14.4 14.6 14.7 14.7 14.6 14.3HD 160934 9.5 10.1 11.2 12.5 13.3 13.9 14.6 14.9 15.0 15.3 15.3 15.2 14.9HD 162283 10.3 11.2 12.2 13.4 14.0 14.6 15.2 15.7 16.1 16.4 16.5 16.5 16.1HD 166181 - 10.8 11.7 13.0 13.7 14.3 15.0 15.4 15.8 16.2 16.5 16.5 16.3HD 167605 9.4 10.5 11.4 12.5 13.3 14.0 14.8 15.1 15.6 15.9 16.1 16.1 15.9HD 187748 - 10.8 11.7 12.9 13.7 14.5 15.3 15.9 16.3 17.0 17.3 17.6 17.4GJ 791.3 9.6 11.0 12.0 13.3 13.8 14.4 15.1 15.6 15.7 16.0 16.1 16.1 15.7HD 197481 - - - 11.0 11.7 12.4 13.5 14.3 14.7 15.5 16.1 16.4 16.3HD 201651 10.1 11.4 12.3 13.3 14.1 14.6 15.3 15.8 16.1 16.4 16.5 16.5 16.3HD 202575 - - 11.4 12.5 13.3 14.0 14.9 15.5 16.1 16.6 16.8 17.0 16.7GJ 4199 10.5 11.2 12.0 13.2 13.8 14.5 15.1 15.6 15.7 16.0 16.1 15.8 15.4HD 206860 - - 12.2 13.3 13.8 14.5 15.2 15.7 16.0 16.5 16.9 17.0 16.7HD 208313 - 11.9 13.0 14.0 14.7 15.2 16.0 16.5 16.7 17.2 17.3 17.3 17.0V383 Lac 10.2 11.0 11.9 13.0 13.6 14.3 14.8 15.2 15.5 15.9 16.1 16.0 15.8HD 213845 - - - 13.3 14.0 14.7 15.7 16.3 16.8 17.2 17.6 18.1 18.0GJ 875.1 9.6 10.5 11.2 12.3 13.1 13.5 14.4 14.9 15.1 15.5 15.6 15.5 15.1GJ 876 - - - 11.0 12.2 12.6 13.7 14.3 15.1 15.8 16.2 16.6 16.6GJ 9809 11.3 12.1 12.8 14.0 14.6 15.0 15.5 15.9 15.9 16.2 16.3 16.1 15.8HD 220140 - - 12.0 13.1 13.9 14.5 15.3 15.8 16.1 16.4 16.6 16.5 16.3HD 221503 - 10.4 11.8 13.2 14.1 14.6 15.3 15.8 16.2 16.7 17.0 17.1 17.0GJ 900 8.9 10.1 10.8 12.4 13.2 13.9 14.9 15.4 15.8 16.1 16.2 16.1 16.0GJ 907.1 8.4 9.0 10.0 11.2 12.1 12.5 13.4 14.0 14.3 14.8 15.0 15.1 14.9Median 9.5 10.5 11.5 12.9 13.6 14.2 15.0 15.5 15.8 16.3 16.5 16.5 16.3Best 11.3 12.1 13.0 14.0 14.7 15.4 16.2 16.8 17.1 17.6 17.8 18.5 18.9 a Magnitude difference in the NIRI CH4-short filter, at a 5 σ level. NIRI CH4-short filter, of a planet of given mass orbiting agiven target. The absolute H -band magnitude of the planetwas first obtained directly from the evolution models ofBaraffe et al. (2003) and converted into an apparent magni-tude, H pl , using the distance of the star. The correspondingmagnitude in the NIRI CH4-short filter was then calculated as m pl = H pl - . (cid:18) f CH4 f H (cid:19) , (2)where f CH4 and f H are the mean flux density of the planet inthe NIRI CH4-short and broad band H filters, respectively;their values were calculated using a synthetic spectrum of ap-propriate effective temperature and surface gravity (Baraffe et al. 2003; Allard et al. 2001) . We recall here (c.f. §3)that the ratio f CH4 f H is typically 1.5–2.5 for giant planets de-pending on their mass and age. The stellar magnitudes in theNIRI CH4-short and broad band H filters were assumed tobe equal, such that the contrast of the planet was obtained as m pl - H ⋆ , where H ⋆ is the H -band apparent magnitude of thetarget star. The 5 σ contrast levels of planets of various massesorbiting a K0 primary of 100 Myr, the typical target of thesurvey, are presented in Figure 4. For a typical target locatedat 22 pc from the Sun, the median detection limits correspond Spectra available at ftp://ftp.ens-lyon.fr/pub/users/CRAL/fallard/ An H -band absolute magnitude of 4.0 was used, this is the mean valueof the K0 stars in the sample. F IG . 5.— Final S/N residual images for three sequences of images to which fiducial point sources have been implanted. The fiducial point sources have beenadded at 5 P.A.’s (0 ◦ , 72 ◦ , 144 ◦ , 216 ◦ , and 288 ◦ ) and 3 angular separations (0.6 ′′ , 1.0 ′′ , and 2.0 ′′ ). For bottom panels (d–f) the intensity of each source was setto the corresponding detection limit (5 σ ) indicated in Table 4, while it was set 0.75 mag brighter (i.e. 10 σ ) for top panels (a–c). Panels (a,d), (b,e), and (c,f) arefor the stars HD 208313, HD 166181, and GJ 507.1, respectively. The bright spot at the upper left corner of panels (a,d) is a real background star. The displayintensity scale is linear from - +
10 for top panels (a–c), and from - + ≥
5) have been circled in white. According to expectations, the detection completeness is roughly 50% for sources whose true intensity is equal tothe 5 σ detection limit. to 10.8 M Jup at 11 AU, 3.9 M Jup at 22 AU, 1.9 M Jup at 44 AU,and 1.4 M Jup at 110 AU.The typical contrast reached by our survey improves on ear-lier surveys (e.g. Lowrance et al. 2005; Masciadri et al. 2005;Chauvin et al. 2006; Biller et al. 2007) by at least 1 mag at1 ′′ , 1.5 mag at 2 ′′ , and ∼ ′′ , our detection limits at thisseparation are similar to those achieved with the SDI deviceat the Very Large Telescope (Biller et al. 2007). The contrastreached by GDPS observations is the highest that has beenachieved to date at separations larger than ∼ ′′ .4.2. Candidate companion detections
To identify candidate companions, the residual images werefirst convolved by a circular aperture of diameter equal to onePSF FWHM, and then converted to signal-to-noise (S/N) im-ages that were visually inspected for point sources at a & σ level. After identification of a point source, its position wasmeasured by fitting a 2D Gaussian function, and its flux wasmeasured in an aperture of diameter equal to one PSF FWHM;both operations were done in the non-convolved residual im-age. The contrast of the point source was then calculated us-ing Eq.(1). More than 300 faint point sources were foundaround 54 targets, 188 of which are found around only 7 starslocated at low galactic latitudes ( b < ◦ ). Up to now, all butsix of the 54 stars with candidates were re-observed at a sub-sequent epoch to verify whether or not the faint point sourcesdetected are co-moving with the target star.All candidate exoplanets observed at two epochs have been S epa r a t i on ( a r cs e c ) P . A . ( deg ) F IG . 6.— Verification of the background nature of the point source detectedaround the young star HD 691. Open diamonds mark the observed separation( top ) and P.A. ( bottom ) of the point source at the two epochs. The solid lineindicates the expected separation and P.A. of a distant background sourceas a function of time. The observations agree very well with the expectedmotion of a background source, indicating that the source is not associatedwith HD 691. confirmed to be background sources by comparing their dis-placement between the two epochs with the expected dis-placement of a distant background source, based on the propermotion and parallax of the target; an example of this verifica-tion is presented in Figure 6. As a reference for future planetsearches, a compilation of all faint point sources identifiedaround our target stars is presented in Table 5.An estimate of the uncertainties on the measured separa-he Gemini Deep Planet Survey 15tions and P.A. was obtained by calculating the mean abso-lute difference between the separation and P.A. measured atthe second epoch and those predicted for this epoch based onthe parallax and proper motion of the target stars. Given thehigh precision on the parallax and proper motion of the target stars, the differences observed are dominated by our measure-ment uncertainties. The mean absolute differences calculatedare taken to represent √ σ sep = 0 . ′′ and σ P . A . = 0 . ◦ are found. TABLE 5P
OINT SOURCES DETECTED
Star Epoch Separation a P.A. b ∆ m c (arcsec) (deg) magHD 166 2005.6482 10.23 82.9 12.60HD 691 2005.6072 2.49 12.1 14.91HD 1405 2004.6409 3.95 254.0 13.98HD 5996 2005.6128 2.98 118.6 12.512005.6128 4.78 71.6 13.232005.6128 5.66 268.9 15.992005.6128 6.95 73.9 15.432005.6128 9.11 280.8 15.72 e e d d d d d d d d d d d d d HD 17382 2004.9740 11.78 130.8 13.16HD 18803 2004.9795 7.61 166.1 17.002004.9795 7.98 208.3 15.682004.9795 10.36 52.8 15.10HD 19994 2005.6648 6.18 187.4 17.622005.6648 6.30 185.3 16.062005.6648 11.64 72.7 17.51HD 283750 2004.8132 7.73 175.8 14.652004.8132 12.72 104.2 13.85HD 30652 2005.6978 2.04 106.3 15.18 d d GJ 182 2004.8459 5.15 220.3 12.802004.8459 7.44 233.7 10.61GJ 234A 2005.8455 3.27 48.8 13.71 d d d d d d GJ 281 2005.2286 5.74 237.0 12.482005.2286 8.80 288.4 12.922006.1158 10.64 224.6 13.43 f GJ 285 2005.2095 8.83 114.3 11.55HD 75332 2005.3107 8.25 141.7 11.49HD 82443 2004.9829 5.27 190.3 11.64 d d d d d TABLE 5 —
Continued
Star Epoch Separation a P.A. b ∆ m c (arcsec) (deg) magHD 90905 2005.2098 5.47 188.2 10.912005.2098 12.41 176.8 13.32HD 92945 2005.3983 9.77 236.2 12.82HD 93528 2005.3271 4.82 332.3 14.27 d GJ 402 2006.1273 12.46 324.0 10.80 f e e HD 108767B 2005.3055 6.72 87.7 12.412005.3055 8.28 100.1 14.622005.3055 10.20 123.9 15.10HD 109085 2005.3986 12.92 256.2 15.80BD+60 1417 2005.2946 2.05 298.4 8.762005.2946 14.08 133.5 12.80 e HD 116956 2005.4067 9.34 17.4 15.05GJ 524.1 2005.2948 7.59 19.7 13.09HD 124106 2005.2975 7.51 124.8 13.852005.2975 9.39 342.2 9.512005.2975 9.60 341.1 8.712005.2975 10.39 287.5 14.652005.2975 11.17 291.7 14.182005.2975 12.06 120.6 15.45HD 130322 2005.4014 7.61 329.8 11.00HD 130948 2005.2922 2.60 103.1 8.56 g g HD 135363 2005.2949 7.50 122.1 10.28HD 139813 2005.4097 6.85 271.3 14.48 d d HD 141272 2005.2977 2.31 12.3 11.442005.2977 4.03 286.9 16.59 e e e e f he Gemini Deep Planet Survey 17 TABLE 5 —
Continued
Star Epoch Separation a P.A. b ∆ m c (arcsec) (deg) mag2005.3007 7.47 75.0 12.532005.3007 7.71 134.8 12.622005.3007 7.94 110.5 15.552005.3007 8.19 198.2 15.442005.3007 8.37 173.6 15.882005.3007 8.60 243.7 14.962005.3007 8.69 132.5 15.012006.7069 9.01 318.9 15.97 f f f e e f f e f e e f f e f e e e e HD 166181 2005.2925 10.38 53.4 14.402005.2925 11.21 195.8 15.042005.2925 13.40 167.6 14.19 e e HD 187748 2005.3965 5.51 325.9 15.762005.3965 7.93 277.1 13.012005.3965 8.02 276.7 12.182005.3965 12.81 114.3 9.742005.3965 13.15 321.5 12.522006.7043 15.02 311.3 14.90 f GJ 791.3 2005.3992 1.98 341.2 12.03 d d d d d d d d d d d d d d d d d d d TABLE 5 —
Continued
Star Epoch Separation a P.A. b ∆ m c (arcsec) (deg) mag2005.3992 9.89 347.0 15.16 d d d d d d d d d d d d d d d d d d d d d d d HD 201651 2005.4867 3.67 201.4 12.482005.4867 8.39 259.1 12.902005.4867 14.53 331.5 13.75 e HD 202575 2005.5386 5.54 28.5 12.412005.5386 12.37 168.0 13.98GJ 4199 2004.6431 9.16 319.9 12.242004.6431 11.76 177.6 10.58HD 206860 2005.6069 3.67 60.0 15.13HD 208313 2005.4868 2.93 30.6 14.782005.4868 6.24 31.1 9.692005.4868 9.45 301.0 14.412005.4868 10.45 137.8 16.50 e e GJ 9809 2006.7019 2.10 240.4 15.30 f f f f f f he Gemini Deep Planet Survey 19 TABLE 5 —
Continued
Star Epoch Separation a P.A. b ∆ m c (arcsec) (deg) mag2006.7019 9.58 84.1 15.30 f f f f e HD 221503 2005.6646 9.02 234.4 15.61 d GJ 900 2004.6458 7.41 76.0 14.202004.6458 12.15 150.6 12.532004.6458 12.41 96.4 9.36GJ 907.1 2005.6837 7.93 296.7 13.68 N OTE . — Target stars around which no point source was detected are omitted from this table. Unless stated otherwise, all point sources listed were confirmed to be backgroundobjects using data from two epochs. a Uncertainty is 0.015 ′′ , see text for detail. b Uncertainty is 0.2 ◦ , see text for detail. c Uncertainties are given in Table 3, see text for detail. d No second epoch data available. e Source undetected in second epoch data. f Source detected in second epoch data only. g Previously known brown dwarf companion (Potter et al. 2002; Goto et al. 2002).
Multiple systems
As mentioned in §2, 16 of the target stars are part of mul-tiple systems. As an orbital solution can potentially be de-termined in a reasonable amount of time for close-separationmultiple systems, the measured properties of the systems ob-served that have a separation below 2 ′′ are presented in Ta-ble 6 as a reference for future studies. Two of the close-separation systems observed were resolved for the first timeby our observations (HD 14802 and HD 166181), and a thirdsystem (HD 213845) was found to be a relatively large sepa-ration ( ∼ ′′ ) binary for which we have found no prior indica-tion in the literature; since HD 213845 is reported to be partof a binary system for the first time in this paper, its propertiesare presented in Table 6 as well. The three multiple systemsobserved for the first time in this work are discussed in moredetail below. HD 14802 — A source 12 ± . ′′ ± . ′′ and P.A.267 . ◦ ± . ◦ (epoch 2005.6348); the large uncertainty onthe flux ratio is due to the peak of the primary star PSF be-ing saturated. Common proper motion of the pair was notverified but the system is likely bound given the brightnessand close separation of the companion. The Hipparcos cata-log (Perryman & ESA 1997) indicates that the proper motionof this star is accelerating and the star is likely part of a binarysystem; an astrometric solution for the system was obtainedby Gontcharov et al. (2000). The estimated period and semi-major axis are 25 yr and 0.5 ′′ , respectively, consistent with the projected separation we have measured. HD 166181 — This star has been known for a long time to bea spectroscopic binary with a period of only 1.8 days (Nadalet al. 1974). More recently, analysis of additional radial ve-locity data has lead Dempsey et al. (1996) to propose that thesystem is in fact triple; a proposition which was confirmed byFekel et al. (2005), who found radial velocity variations as-cribable to a third component with an orbit of period 5.7 yearand eccentricity 0.765. Further, by reanalyzing Hipparcosdata in light of this new component, these authors have founda new astrometric solution for the system, leading to revisedvalues of parallax and proper motion (see Table 1) and to a de-termination of the orbital inclination of the long-period com-panion. Based on their complete solution, they estimate thesemi-major axis of the outer companion at 0.077 ′′ (2.5 AU)and its mass at 0.79 M ⊙ . Our observations have resolved thelong-period companion of this triple system. In 2005.2926,the companion was located at a separation of 0 . ′′ ± . ′′ and P.A. of 16 . ◦ ± . ◦ , and in 2006.7124, it was located at aseparation of 0 . ′′ ± . ′′ and P.A. of 51 . ◦ ± . ◦ . Theevolution of the separation and P.A. of this source between thetwo epochs is far from that expected for an unrelated back-ground source and is in very good agreement with the orbitalmotion expected based on the astrometric solution of Fekelet al. (2005) (see Fig. 7), confirming that the source observedis HD 166181B. The flux ratio of the component Aab to com-ponent B is ∼ ∼ TABLE 6P
ROPERTIES OF CLOSE SEPARATION MULTIPLE SYSTEMS
Name Epoch Separation P.A. Brightness 1 st spatially resolved( ′′ ) ( ◦ ) ratio a observationHD 14802AB 2005.6348 0 . ± .
005 267 . ± . ± . ± .
002 44 . ± . . ± . . ± .
004 355 . ± . . ± . . ± .
002 172 . ± . . ± . . ± .
002 171 . ± . . ± .
002 173 . ± . . ± . . ± .
002 129 . ± . . ± . b . ± .
002 131 . ± . . ± .
002 268 . ± . . ± . . ± .
002 271 . ± . . ± .
005 16 . ± . . ± . . ± .
003 51 . ± . . ± .
002 46 . ± . . ± . . ± .
03 129 . ± . ∼ c This work2006.5108 6 . ± .
03 129 . ± . . ± .
002 334 . ± . . ± . . ± .
002 338 . ± . . ± .
002 344 . ± . . ± . . ± .
002 345 . ± . . ± .
002 213 . ± . . ± .
05 Rossiter (1955)2006.5411 0 . ± .
002 212 . ± . a Brightness of the primary over that of the companion, in the NIRI CH4-short filter. b Independently resolved by Biller et al. (2007), evidence for co-motion is reported in this work for the first time. c The peak of both the primary and companion is saturated in the data; the ratio quoted is an estimate based on the comparison of radial profiles. S epa r a t i on ( a r cs e c ) P . A . ( deg ) F IG . 7.— Verification of the physical association of the point source de-tected around HD 166181. Open diamonds mark the observed separation( top ) and P.A. ( bottom ) of the point source at the two epochs. The solid lineindicates the expected separation and P.A. of a distant background source asa function of time. The predicted separation and P.A. of HD 166181B basedon the astrometric solution of Fekel et al. (2005) are shown as dashed lines ,with uncertainties indicated by the shaded areas. HD 213845 — A bright source is visible in our data at aseparation of 6 . ′′ ± . ′′ and P.A. of 129 . ◦ ± . ◦ fromHD 213845 (epoch 2005.6453). This source did not changeseparation nor P.A. between our 2005 and 2006 observations(see Figure 8), indicating that it is bound to HD 213845. Thecompanion is only visible in our saturated data as its separa-tion exceeds the field of view of the sub-array used for theunsaturated observations. Further, being relatively bright, thepeak of the companion’s PSF is saturated in all our data, mak-ing it very difficult to estimate its flux ratio to the primaryand explaining the larger uncertainty on the separation andP.A. quoted above. We have nevertheless estimated that thecompanion is ∼
125 times fainter than its primary based on acomparison of their radial intensity profiles at radii where thedata are in the linear regime of the detector. The companionwas possibly detected by 2MASS, but its measured position S epa r a t i on ( a r cs e c ) P . A . ( deg ) F IG . 8.— Same as Figure 7 for HD 213845 and photometry in the 2MASS point source catalog (PSC) areaffected by confusion due to the nearby primary. Neverthe-less, the relative position of this source in the 2MASS PSC,separation of 5.55 ′′ and P.A. of 128 ◦ , is consistent with thestar being gravitationally bound to HD 213845 as, were itnot a bound companion, its separation should have changedby ∼ ′′ between the 2MASS observations and our first epochobservations. Although the separation of this binary system iswell above the resolution limit of seeing-limited observations,we have found no prior indication of binarity in the literature. ANALYSIS AND DISCUSSIONThe detection limits determined in §4.1 can be used to cal-culate an upper limit to the fraction of stars that have compan-ions of mass and semi-major axis inside some given intervals.The analysis presented in this section is largely guided by thework of Brandeker et al. (2006); Carson et al. (2006); Allenet al. (2005); and Sivia (1996). The statistical formalism forthe analysis is presented first and various applications to ourdata are presented afterward.he Gemini Deep Planet Survey 215.1.
Statistical formalism
Consider the observation of N stars enumerated by j =1 . . . N . Let f be the fraction of stars that have at leastone companion of mass and semi-major axis in the intervals[ m min , m max ] and [ a min , a max ], respectively, and p j the proba-bility that such a companion around star j , if indeed it wasthere, would be detected given the detection limits of the ob-servations. The probability of detecting such a companionaround star j is f p j , and the probability of not detecting acompanion around this star is simply 1 - f p j . If the set { d j } denotes the detections made by the observations, such that d j equals 1 if a companion is detected around star j or else equals0, then the probability that the observed outcome would oc-cur, also called the likelihood of the data given f , is givenby L ( { d j }| f ) = N Y j =1 (cid:0) - f p j (cid:1) ( - d j ) (cid:0) f p j (cid:1) d j . (3)According to Bayes’ theorem, from the a priori probabilitydensity p ( f ), or prior distribution, and the likelihood function L , one may calculate p ( f |{ d j } ), the probability density up-dated in light of the data, or posterior distribution: p ( f |{ d j } ) = L ( { d j }| f ) p ( f ) R L ( { d j }| f ) p ( f )d f . (4)In this study, since we have no prior knowledge about f , weuse the most ignorant prior distribution p ( f ) = 1.The posterior distribution p ( f |{ d j } ) can be used to deter-mine a credible interval (CI) for f , bounded by f min and f max ,for a given level of credibility α . For a case where there is nodetection, as is the case with our survey, then clearly f min = 0,and the upper bound of the CI is found by solving α = Z f max p ( f |{ d j } )d f . (5)For a case where there are some detections, an equal-tail CI isfound by solving1 - α Z f min p ( f |{ d j } )d f and 1 - α Z f max p ( f |{ d j } )d f . (6)In this work a value of α = 0 .
95 was chosen.The determination of the p j ’s is a critical step of this anal-ysis; their value depends on the detection limits of the obser-vations, on the ages and distances of the systems, and on themass, semi-major axis, and orbital eccentricity distributionsof the companions. In calculating the p j ’s it is also impor-tant to account properly for orbital inclination and phase asthese affect significantly the distribution of projected separa-tions for an orbit of given semi-major axis. In this work, the p j ’s were calculated using a Monte Carlo approach. The massand semi-major axis intervals, [ m min , m max ] and [ a min , a max ],were first selected. Then for each target star, 10000 planetswere generated by sampling randomly, for each planet, themass, semi-major axis, orbital eccentricity, orbital separationprojection factor, age of the system, and underlying residualnoise in the image. The mass and semi-major axis distribu-tions are left arbitrary for the moment; different possibilitieswill be explored later. For all of our calculations, the orbitaleccentricity distribution was assumed to be that of the radialvelocity exoplanets sample, which was approximated by aGaussian function of mean 0.25, standard deviation 0.19, and with 0 ≤ e ≤ . σ detection limit, while the sig-nal of some “a priori undetectable” planets will be boostedabove the detection limit such that the appropriate detectioncompleteness will result for planets of various true intensi-ties (see §4.1). Finally, given the sample of planets assignedto target j , the probability p j was calculated as the fractionof planets lying above the corresponding 5 σ detection limits(c.f. Table 4).The above determination of the p j ’s yields a CI for f thatis a function of the assumptions made on the mass and semi-major axis distributions. For a case where there is no detec-tion, it is also possible to obtain a more conservative estimateof f max that is valid for any distributions of mass and semi-major axis. The procedure used to do this is identical to thatdescribed above except for the following. Rather than popu-lating the whole intervals of mass and semi-major axis con-sidered, all planets are assigned a mass and semi-major axisprecisely equal to m min and a min , respectively. Because moremassive or more distant planets are easier to detect, the val-ues of p j ’s calculated in this manner constitute lower limitsto the values that would be obtained by populating the wholeintervals assuming any specific distributions; accordingly, theresulting value of f max constitutes an upper limit. This ap-proach is perfectly legitimate as long as a max is chosen suchthat the values of p j ’s for any a in [ a min , a max ] are at least aslarge as those for a min .It is possible to derive a simple analytic expression for f max for a case where there is no detection; this expression may beuseful to estimate what the results of an ongoing survey willbe or scale actual results for different values of N or detec-tion probabilities. This expression may be obtained by firstreplacing each p j by the average detection probability h p j i inEq. (3), and then recognizing that the likelihood function canbe approximated by e - N f h p j i . This leads to f max ≈ - ln (1 - α ) N h p j i . (7)This approximation, valid for N h p j i ≫
1, is equivalent to us-ing Poisson statistics rather than Binomial statistics for thepresence of companions (c.f. Eq. 3–7 of Carson et al. 2006).5.2. f max for arbitrary mass and semi-major axisdistributions As a first analysis of the survey results, we present estimatesof f max that are independent of the mass and semi-major axisdistributions for m min =0.5, 1, 2, 3, 4, 5, 7.5, 10, and 13 M Jup ,and for all a min between 10 and 500 AU; these estimates were2 Lafrenière et al.
10 20 30 40 50 100 200 300 400 500Semi−major axis (AU)0102030405060708090100 M ean p r obab ili t y o f de t e c t i on ( % ) Jup F IG . 9.— Mean probability of detection of a planet of given mass as afunction of the semi-major axis of its orbit; the curves are labeled by themass of the planet, in M Jup . The mean is obtained over all targets of thesurvey.
10 20 30 40 50 100 200 300 400 500Semi−major axis (AU)0.00.10.20.30.40.50.60.70.80.91.0 P l ane t f r equen cy uppe r li m i t Jup F IG . 10.— Upper limits, with a credibility of 95%, on the fraction ofstars harboring at least one companion of mass in the range [ m min , M Jup and orbit of semi-major axis in various ranges. The minimum mass, m min , isindicated on each curve. For any interval, [ a min , a max ] AU, of semi-major axisselected, the correct value of f max to read from the graph is the maximum ofthe curve within that interval. The curves shown in this graph are conservativeupper limits that are valid for any distributions of mass and semi-major axis.The dotted line indicates the minimum upper limit that one could derive fromobservation of 79 stars if the probability of detection of a planet was 100%irrespective of its age, mass, and orbital separation. calculated according to the last procedure described above.For this analysis, and those in the next section, we have notconsidered the 6 stars with candidates for which second epochobservations are missing. The results obtained in this sectionare valid for any m max up to ∼ M Jup as no companion witha mass below this value was detected. Planet detection prob-abilities for each star are indicated in Table 7 for a small se-lection of masses and semi-major axes, while the mean planetdetection probabilities, i.e. the average of the p j ’s over all j ’s, are shown in Figure 9 as a continuous function of semi-major axis and for a larger selection of masses. The peaksensitivity of the survey occurs for semi-major axes between50 and 200 AU; the peak values are ∼
45% and ∼
68% for 2and 5 M Jup , respectively. The survey is particularly sensitiveto brown dwarfs ( m & M Jup ), with a detection probabilityabove 75% between 35 and 200 AU. A decline in sensitiv-ity occurs at a separation of ∼
200 AU; this is consistent withthe field of view of the observations ( ∼ ′′ radius) and mean The previously known 40–65 M Jup binary brown dwarf companion lo-cated 2.6 ′′ from HD 130948 (Potter et al. 2002; Goto et al. 2002) is detectedin our data.
10 20 30 40 50 100 200 300 400 500Semi−major axis (AU)0.00.10.20.30.40.50.6 P l ane t f r equen cy uppe r li m i t F IG . 11.— Upper limits, with a credibility of 95%, on the fraction of starsharboring at least one planet of mass in the range [0 . , M Jup , assumingd n / d m ∝ m β , and semi-major axis in various ranges. The values of β are - dot-dashed line ), - . solid line ), and 0 ( dashed line ). For any interval,[ a min , a max ] AU, of semi-major axis selected, the correct value of f max to readfrom the graph is the maximum of the curve within that interval. The 67%credibility curve for β = - . dotted line ). distance of the targets (22 pc).The results for f max are shown in Figure 10. For a semi-major axis interval lower bound of 50 AU, the 95% credi-ble planet frequency upper limits are 0.28 for 1–13 M Jup and50–225 AU, 0.12 for 2–13 M Jup and 50–295 AU, and 0.057for 5–13 M Jup and 50–185 AU. For a semi-major axis lowerbound of 25 AU, the upper limits are 0.23 for 2–13 M Jup and25–420 AU and 0.09 for 5–13 M Jup and 25–305 AU. For com-pleteness, the exercise was repeated for circular orbits and fora uniform distribution of eccentricity (between 0 and 1), andthe results obtained were very similar to those shown in Fig-ure 10.The results also indicate that no more than 0.056 of starshave low-mass brown dwarf companions (13 < m / M Jup < < m / M Jup <
75) over the same range of semi-major axis, the brown dwarf companion to HD 130948 (Pot-ter et al. 2002; Goto et al. 2002) must be taken into accountexplicitly. This analysis must be carried out with care asthe semi-major axis of this companion could be significantlydifferent from its measured projected physical separation of47 AU. It is possible to account for this uncertainty by calcu-lating the probability distribution of the real semi-major axisof the brown dwarf companion using a Monte Carlo approachsimilar to the one presented above for the calculation of the p j ’s. Basically, the projected separation of the companion isfixed at s = 47 AU and its orbital eccentricity and orbital pro-jection factor are sampled randomly 10 times, as describedabove. The de-projected semi-major axis is then calculatedfor each random trial and its normalized distribution over alltrials is obtained. As the projection factor can never be largerthan (1 + e max ), where e max is the maximum eccentricity al-lowed, the semi-major axis probability distribution is equalto zero below s / (1 + e max ); the distribution extends to infinityfor higher values. Applied to the current case, this calculationindicates that at a 95% credible interval for the semi-majoraxis of the binary brown dwarf companion to HD 130948 is26–157 AU. We thus posit that our observations have resultedin one detection in the semi-major axis interval 25–200 AUand mass interval 13–75 M Jup ; then using the procedure de-scribed in the previous section and Eq. (6), the 95% credi-ble interval for the frequency of stars with at least one brownhe Gemini Deep Planet Survey 23 a min = 10 AU a min = 25 AU a min = 50 AU
10 20 30 40 50 100 200 300 400 500 a max (AU)0.000.050.100.150.20 P l ane t f r equen cy uppe r li m i t F IG . 12.— Upper limits, with a credibility of 95% ( top panel ) or 67% ( bottom panel ), on the fraction of stars harboring at least one giant planet of mass inthe range [0 . , M Jup , assuming d n / d m ∝ m - . , and orbit of semi-major axis in the range [ a min , a max ] AU, assuming d n / d a ∝ a γ . The abscissa indicates theupper bounds ( a max ) of the semi-major axis intervals, while the lower bounds ( a min ) are 10 AU ( solid lines ), 25 AU ( dotted lines ), and 50 AU ( dashed lines ). Thetop, middle, and bottom curves in each set of three curves are for γ = -
1, 0, and 1, respectively. dwarf companion in the semi-major axis interval 25–250 AUis 0 . + . - . . This result is consistent with the upper limit of0.12 (95% credibility) reported by Carson et al. (2006) for the25–100 AU semi-major axis interval and also with the frac-tion of 0 . + . - . (95% credibility) reported by Metchev &Hillenbrand (2004) for the range 30–1600 AU. For smallersemi-major axes, our results indicate that, with a credibilityof 95%, the fraction of stars with at least one brown dwarfcompanion in the range 10–25 AU is less than 0.20, and lessthan 0.10 for the range 15–25 AU.5.3. f max for specific mass and semi-major axis distributions In this section we derive first an upper limit to the fractionof stars harboring at least one planet in the single mass inter-val [0 . , M Jup , assuming that the mass distribution followsd n / d m ∝ m - . . The mass distribution adopted is based on astatistical analysis of the RV results that properly accounts forthe detection sensitivity reached for each star (A. Cumming etal. 2007, in preparation) and is formally valid only for planetswith semi-major axis below ∼ n / d m ∝ m - . (Butler et al. 2006). For this calculation thewhole mass interval is populated according to the distributionstated, but all planets are assigned a value a min for the semi-major axis, so as to make the results independent of its dis-tribution. The calculation was made for all a min between 10and 500 AU. The results are shown in Figure 11. With a cred-ibility of 95%, the fraction of stars having at least one planetof mass in the range [0 . , M Jup and semi-major axis in [10 , , , n / d m ∝ m β , with β = 0 and -
2, are presentedalso in Figure 11. As expected, a smaller β leads to larger val-ues of f max because a larger fraction of planets have a smallermass, while a larger value of β has the opposite effect.Next we calculate upper limits for the same mass intervalby assuming further that the distribution of semi-major axesfollows d n / d a ∝ a γ , for γ = - ,
0, and 1. This range ofpower-law index includes the value of γ = - .
67 found by A.Cumming et al. (2007, in preparation) for the RV exoplanetssample within the range 0.03–3 AU. We have done the cal-culations for a min =10, 25, and 50 AU, and for all a max in therange [ a min + , γ = -
1, the 95% credible upper limitsto the fraction of stars with at least one planet of mass in therange [0 . , M Jup are 0.28 for the semi-major axis range10–25 AU, 0.18 for 10–50 AU, 0.13 for 25–50 AU, 0.11 for25–100 AU, and 0.093 for 50–250 AU. Slightly smaller val-ues of f max are found for larger values of γ , as such indiceswould place more planets at larger separations where theywould have been easier to detect with our observations. Forthe larger values of a min , the value of γ has very little effecton the upper limit found as, irrespective of the value of γ ,the majority of planets are located at separations for whichthe sensitivity of the observations is high. Overall, the weak4 Lafrenière et al. F IG . 13.— Detection limits (5 σ , solid line ) and synthetic population ofplanets ( dots ) for the star HD 166181. A planet mass distribution following dn / dm ∝ m - . inside 0.5–13 M Jup and a semi-major axis distribution follow-ing dn / da ∝ a - inside 10–300 AU were used. For this particular example,the planet detection probability p j is 30%. dependence of f max on γ implies that the semi-major axis dis-tribution (i.e. γ ) cannot be constrained from our results. As one may worry that the population of planets around Mdwarfs is different from that around earlier-type stars, becauseof smaller disk masses for example, we derive an estimate of f max by excluding the M dwarfs from the statistical analysis.This estimate is obtained using Eq. (7) and the values of thelast three columns of Table 7; it is thus valid for β = - . γ = -
1. Excluding M dwarfs from the sample leaves 64 starsand results in average detection probabilities h p j i of 0.070,0.229, and 0.385 for 10–25 AU, 25–50 AU, and 50–250 AU,respectively. The corresponding 95% credible upper limits tothe fraction of stars with planets are then 0.67, 0.20, and 0.12.The effect is quite significant at the smallest orbital separa-tions, where M dwarfs provide good sensitivities due to theirsmaller luminosity, smaller average distance, and younger av-erage age. Similarly, the population of planets in stellar mul-tiple systems may be different from that in single systems.Excluding multiples from the sample also leaves 64 stars andyields values of h p j i of 0.139, 0.299, and 0.389 for 10–25 AU,25–50 AU, and 50–250 AU. The corresponding upper limitsto the fraction of stars with planets are 0.34, 0.16, and 0.12;the effect is thus rather small in this case. TABLE 7P
LANET DETECTION PROBABILITY M Jup M Jup M Jup , β = - . γ = - he Gemini Deep Planet Survey 25 TABLE 7 —
Continued M Jup M Jup M Jup , β = - . γ = - SUMMARY AND CONCLUSIONIn this paper, we have presented the results of the GeminiDeep Planet Survey, a near-infrared adaptive optics search forgiant planets on orbits of 10–300 AU around nearby youngstars. The use of angular differential imaging at the GeminiNorth telescope has enabled us to reach the best sensitivitiesto date for detecting giant exoplanets with projected separa-tions above ∼ ′′ . The typical detection limits (5 σ ) reachedby the survey, in magnitude difference between an off-axispoint source and the central star, are 9.5 at 0.5 ′′ , 12.9 at 1 ′′ ,15 at 2 ′′ , and 16.5 at 5 ′′ , sufficient to detect planets more mas-sive than 2 M Jup with a projected separation of 40–200 AUaround a typical target star. More than 300 faint point sourceshave been detected around 54 of the 85 stars observed, butobservations at a second epoch have revealed changes in sep-aration and P.A. of these point sources relative to the targetstars that are all consistent with those expected from unre-lated background objects. The observations made as part ofthis survey have resolved the stars HD 14802, HD 166181,and HD 213845 into binaries for the first time.We have presented a statistical analysis of the survey resultsto derive upper limits to the fraction of stars having planetarycompanions. This analysis indicates that the 95% credible up-per limit to the fraction of stars harboring at least one planetmore massive than 2 M Jup with an orbit of semi-major axis inthe range 25–420 AU or 50–295 AU is 0.23 or 0.12, respec-tively, independently of the mass and semi-major axis distri-butions of the planets; for planets more massive than 5 M Jup ,the upper limits are 0.09 for 25-305 AU and 0.057 for 50–185 AU. It was also found that less than 0.056 of stars havelow-mass brown dwarf companions (13 < m / M Jup <
40) be-tween 25 and 250 AU (see Figure 10); this upper limit is set by the sample size only as the sensitivity of the observationsto brown dwarfs is very good. Considering the whole browndwarf mass range, the 95% credible interval for the frequencyof stars with at least one brown dwarf companion in the semi-major axis interval 25–250 AU is 0 . + . - . . Assuming amass distribution following d n / d m ∝ m - . , the results indi-cate that with a credibility of 95% the fraction of stars havingat least one planet of mass in the range 0.5–13 M Jup and semi-major axis in the range 25–325 AU is less than 0.17, and lessthan 0.10 for the range 50–220 AU. Assuming further a semi-major axis distribution following d n / d a ∝ a - , the upper lim-its to the fraction of stars with planets are 0.28 for the range10–25 AU, 0.13 for 25–50 AU, and 0.093 for 50–250 AU.The work presented in this paper constitutes a first step to-ward the detection of the population of “outer” giant planetsaround other stars. Such a study, which is complementary toRV searches in terms of orbital separation, is necessary to im-prove our understanding of the various mechanisms that couldgenerate planets on orbits of tens to hundreds of AU, such asin situ formation triggered by collisions of stars with proto-planetary disks or orbital migration induced by gravitationalscattering in multiple planet systems. While the upper limitswe have found rule out an important increase in the populationof planets at large separations compared to the known popu-lation of planets below 3 AU, our sample size and the sensi-tivities we haved reached are insufficient to tell if the abovemechanisms operate at all, and a fortiori which one is domi-nant. Future searches reaching better sensitivities and targetedat a larger sample of stars will be necessary to answer thesequestions.Considerable efforts are currently deployed by major ob-servatories to develop instruments dedicated to the search of6 Lafrenière et al.giant exoplanets around nearby stars. The Gemini Planet Im-ager (GPI, Gemini Telescope, Macintosh et al. 2006) andthe Spectro-Polarimetric High-contrast Exoplanet Researchinstrument (SPHERE, Very Large Telescope, Dohlen et al.2006) are good examples; they should see their first light inaround 2010. These complex instruments will ally an extremeAO system to correct atmospheric wavefront errors to un-precedented levels of accuracy, a calibration system to correctinstrumental quasi-static aberrations, a coronagraph to sup-press the coherent on-axis stellar light, and differential imag-ing capabilities enabled by either multi-channel cameras or in-tegral field spectrographs. The expected performance of theseinstruments, e.g. a contrast better than 17.5 mag at a sepa-ration of 0.5 ′′ for GPI (Macintosh et al. 2006), should allowdetection of planets of 1 M Jup aged less than 100–200 Myr atseparations of 5–50 AU, significantly improving on the workpresented here. These efforts should uncover the populationof outer giant planets, if they exist, or place sufficient con-straints on their existence to rule out the mechanisms thatcould generate them. In less than a decade the James WebbSpace Telescope will allow similar studies to be done for rela-tively nearby M-type primaries, which are too faint for operat-ing the wavefront sensor of extreme adaptive optics systems.Given all of the projects that should unfold in the next fewyears, the coming decade promises to be extremely exciting for exoplanet science.We are grateful to the referee whose thorough review andexcellent suggestions have improved the quality of this pa-per significantly. The authors would like to thank the Geministaff for carrying out all the observations. This project wasmade possible through the support and generous allocationof observing time from the Canadian, US, UK, and Geministaff time allocation committees. This work was supported inpart through grants from the Natural Sciences and Engineer-ing Research Council, Canada, from the Fonds Québécois dela Recherche sur la Nature et les Technologies, and from theFaculté des Études Supérieures de l’Université de Montréal.This research was performed in part under the auspices ofthe US Department of Energy by the University of California,Lawrence Livermore National Laboratory under contract W-7405-ENG-48, and also supported in part by the National Sci-ence Foundation Science and Technology Center for AdaptiveOptics, managed by the University of California at Santa Cruzunder cooperative agreement AST 98-76783. This researchhas made use of the SIMBAD database, operated at Centrede Données astronomiques de Strasbourg (CDS), Strasbourg,France. This research has made use of the VizieR catalog ser-vice (Ochsenbein et al. 2000), hosted by the CDS.
REFERENCESAlibert, Y., Mordasini, C., Benz, W., & Winisdoerffer, C. 2005, A&A, 434,343Allard, F., Hauschildt, P. H., Alexander, D. R., Tamanai, A., & Schweitzer,A. 2001, ApJ, 556, 357Allen, P. R., Koerner, D. W., Reid, I. N., & Trilling, D. E. 2005, ApJ, 625,385Alonso, R. et al. 2004, ApJ, 613, L153Arribas, S., Mediavilla, E., & Fuensalida, J. J. 1998, ApJ, 505, L43+Augereau, J. C., Lagrange, A. M., Mouillet, D., & Ménard, F. 1999, A&A,350, L51Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 2002, A&A, 382, 563Baraffe, I., Chabrier, G., Barman, T. S., Allard, F., & Hauschildt, P. H. 2003,A&A, 402, 701Barrado y Navascués, D., Stauffer, J. R., Song, I., & Caillault, J.-P. 1999,ApJ, 520, L123Bate, M. R., & Bonnell, I. A. 2005, MNRAS, 356, 1201Beaulieu, J.-P. et al. 2006, Nature, 439, 437Biller, B. A., Close, L., Lenzen, R., Brandner, W., McCarthy, D. W., Nielsen,E., & Hartung, M. 2004, in Proc. SPIE, Vol. 5490, Advancements inAdaptive Optics, ed. D. Bonaccini Calia, B. L. Ellerbroek, & R. Ragazzoni,389–397Biller, B. A. et al. 2007, ApJS, in press. (astro-ph/0705.0066)Bond, I. A. et al. 2004, ApJ, 606, L155Boss, A. P. 1997, Science, 276, 1836—. 2000, ApJ, 536, L101—. 2001, ApJ, 563, 367—. 2003, ApJ, 599, 577—. 2006, ApJ, 637, L137Brandeker, A., Jayawardhana, R., Khavari, P., Haisch, Jr., K. E., & Mardones,D. 2006, ApJ, 652, 1572Burrows, A., Marley, M., Hubbard, W. B., Lunine, J. I., Guillot, T., Saumon,D., Freedman, R., Sudarsky, D., & Sharp, C. 1997, ApJ, 491, 856Butler, R. P. et al. 2006, ApJ, 646, 505Cameron, A. C. et al. 2007, MNRAS, 375, 951Carpenter, J. M., Wolf, S., Schreyer, K., Launhardt, R., & Henning, T. 2005,AJ, 129, 1049Carson, J. C., Eikenberry, S. S., Smith, J. J., & Cordes, J. M. 2006, AJ, 132,1146Cayrel de Strobel, G., Soubiran, C., Friel, E. D., Ralite, N., & Francois, P.1997, A&AS, 124, 299Cayrel de Strobel, G., Soubiran, C., & Ralite, N. 2001, A&A, 373, 159Charbonneau, D., Brown, T. M., Latham, D. W., & Mayor, M. 2000, ApJ,529, L45Chatterjee, S., Ford, E. B., & Rasio, F. A. 2007, ArXiv Astrophysics e-prints,astro-ph/0703166Chauvin, G., Lagrange, A.-M., Dumas, C., Zuckerman, B., Mouillet, D.,Song, I., Beuzit, J.-L., & Lowrance, P. 2004, A&A, 425, L29—. 2005a, A&A, 438, L25Chauvin, G., Lagrange, A.-M., Udry, S., Fusco, T., Galland, F., Naef, D.,Beuzit, J.-L., & Mayor, M. 2006, A&A, 456, 1165Chauvin, G. et al. 2005b, A&A, 438, L29 Clem, J. L., VandenBerg, D. A., Grundahl, F., & Bell, R. A. 2004, AJ, 127,1227Close, L. M. et al. 2007, ApJ, 660, 1492Deller, A. T., & Maddison, S. T. 2005, ApJ, 625, 398Dempsey, R. C., Neff, J. E., O’Neal, D., & Olah, K. 1996, AJ, 111, 1356Dohlen, K. et al. 2006, in Proc. SPIE, Vol. 6269, Ground-based and AirborneInstrumentation for Astronomy, ed. I. S. McLean & M. Iye, 62690QDonahue, R. A. 1993, PhD thesis, AA(New Mexico State Univ., UniversityPark.)Favata, F., Barbera, M., Micela, G., & Sciortino, S. 1995, A&A, 295, 147Favata, F., Micela, G., Sciortino, S., & D’Antona, F. 1998, A&A, 335, 218Fekel, F. C., Barlow, D. J., Scarfe, C. D., Jancart, S., & Pourbaix, D. 2005,AJ, 129, 1001Fernandez, J. A., & Ip, W.-H. 1984, Icarus, 58, 109Fuhrmann, K. 2004, Astronomische Nachrichten, 325, 3Gaidos, E. J. 1998, PASP, 110, 1259Gaidos, E. J., Henry, G. W., & Henry, S. M. 2000, AJ, 120, 1006Gatewood, G., Coban, L., & Han, I. 2003, AJ, 125, 1530Gerbaldi, M., Faraggiana, R., & Balin, N. 2001, A&A, 379, 162Gontcharov, G. A., Andronova, A. A., & Titov, O. A. 2000, A&A, 355, 1164Goto, M. et al. 2002, ApJ, 567, L59Gould, A. et al. 2006, ApJ, 644, L37Gray, R. O., Corbally, C. J., Garrison, R. F., McFadden, M. T., & Robinson,P. E. 2003, AJ, 126, 2048Greaves, J. S. et al. 1998, ApJ, 506, L133Hahn, J. M., & Malhotra, R. 2005, AJ, 130, 2392Haisch, Jr., K. E., Lada, E. A., & Lada, C. J. 2001, ApJ, 553, L153Heinze, A., Hinz, P., Sivanandam, S., & Meyer, M. 2006, in Bulletin ofthe American Astronomical Society, Vol. 38, Bulletin of the AmericanAstronomical Society, 1212Henry, T. J., Soderblom, D. R., Donahue, R. A., & Baliunas, S. L. 1996, AJ,111, 439Herriot, G. et al. 2000, in Proc. SPIE, Vol. 4007, Adaptive Optical SystemsTechnology, ed. P. L. Wizinowich, 115–125Hinkley, S. et al. 2007, ApJ, 654, 633Hodapp, K. W. et al. 2003, PASP, 115, 1388Høg, E. et al. 1997, A&A, 323, L57Høg, E. et al. 2000, A&A, 355, L27Holland, W. S. et al. 1998, Nature, 392, 788Hormuth, F., Brandner, W., Hippler, S., Janson, M., & Henning, T. 2007,A&A, 463, 707Huensch, M., Schmitt, J. H. M. M., & Voges, W. 1998, A&AS, 132, 155Hünsch, M., Schmitt, J. H. M. M., Sterzik, M. F., & Voges, W. 1999, A&AS,135, 319Ida, S., & Lin, D. N. C. 2004, ApJ, 604, 388Itoh, Y. et al. 2005, ApJ, 620, 984Jayawardhana, R., Coffey, J., Scholz, A., Brandeker, A., & van Kerkwijk,M. H. 2006, ApJ, 648, 1206Kalas, P., Graham, J. R., & Clampin, M. 2005, Nature, 435, 1067Kasper, M., Apai, D., Janson, M., & Brandner, W. 2007, ArXiv e-prints,astro-ph/0706.0095 he Gemini Deep Planet Survey 27he Gemini Deep Planet Survey 27