Deep infrared imaging of close companions to austral A- and F-type stars
David Ehrenreich, Anne-Marie Lagrange, Guillaume Montagnier, Gaël Chauvin, Franck Galland, Jean-Luc Beuzit, Julien Rameau
AAstronomy & Astrophysics manuscript no. preprint October 25, 2018
Deep infrared imaging of close companionsto austral A- and F-type stars (cid:63),(cid:63)(cid:63)
D. Ehrenreich , A.-M. Lagrange , G. Montagnier , G. Chauvin , F. Galland , J.-L. Beuzit & J. Rameau Laboratoire d’astrophysique de Grenoble, Universit´e Joseph Fourier, CNRS (UMR 5571), BP 53, 38041 Grenoble cedex 9, Francee-mail: [email protected] European Southern Observatory, Alonso de Cordova 3107, Vitacura Casilla 19001, Santiago 19, Chile
ABSTRACT
The search for substellar companions around stars with di ff erent masses along the main sequence is critical to understand the di ff erentprocesses leading to the formation of low-mass stars, brown dwarfs, and planets. In particular, the existence of a large population oflow-mass stars and brown dwarfs physically bound to early-type main-sequence stars could imply that the massive planets recentlyimaged at wide separations (10–100 au ) around A-type stars are disc-born objects in the low-mass tail of the binary distribution andare thus formed via gravitational instability rather than by core accretion. Our aim is to characterize the environment of early-typemain-sequence stars by detecting substellar companions between 10 and 500 au . The sample stars are also surveyed with radialvelocimetry, providing a way to determine the impact of the imaged companions on the presence of planets at < ∼ au . High contrastand high angular resolution near-infrared images of a sample of 38 southern A- and F-type stars have been obtained between 2005and 2009 with the instruments NaCo on the Very Large Telescope and PUEO on the Canada-France-Hawai‘i Telescope. Direct andsaturated imaging were used in the J to K s bands to probe the faint circumstellar environments with contrasts of ∼ × − to 10 − at separations of 0 . (cid:48)(cid:48) (cid:48)(cid:48) , respectively. Using coronagraphic imaging, we achieved contrasts between 10 − and 10 − at separations > (cid:48)(cid:48) . Multi-epoch observations were performed to discriminate comoving companions from background contaminants. This surveyis sensitive to companions of A and F stars in the brown dwarf to low-mass star mass regime. About 41 companion candidateswere imaged around 23 stars. Follow-up observations for 83% of these stars allowed us to identify a large number of backgroundcontaminants. We report the detection of 7 low-mass stars with masses between 0.1 and 0.8 M (cid:12) in 6 multiple systems: the discoveryof a M2 companion around the A5V star HD 14943 and the detection of HD 41742B around the F4V star HD 41742 in a quadruplesystem; we resolve the known companion of the F6.5V star HD 49095 as a short-period binary system composed by 2 M / L dwarfs.We also resolve the companions to the astrometric binaries ι Crt (F6.5V) and 26 Oph (F3V), and identify a M3 / M4 companion to theF4V star o Gru, associated with a X-ray source. The global multiplicity fraction measured in our sample of A and F stars is ≥ Key words.
Planetary systems – Stars: early type – Stars: binaries (including multiple): close – Surveys – Astrometry –Instrumentation: adaptive optics
1. Introduction
Stars do not form alone, as they are often found in multiple sys-tems, commonly consisting of two or more stellar components,brown dwarfs, or planets. The various properties of these sys-tems in terms of mass ratios and separations call for di ff erentformation mechanisms. At the lower end of the mass distribu-tion, more than 400 extrasolar planets have been detected, mostof them ( ∼ < ∼ au , which correspond to the largest revolution periodsprobed by radial velocimetry today, are preferentially formingthrough rapid accretion of gas on pre-existing rocky cores, mas- Send o ff print requests to : D. Ehrenreich (cid:63) Based on observations made with ESO Telescopes at the ParanalObservatory under programme IDs 076.C-0270, 081.C-0653, and083.C-0151, and on observations obtained at the Canada-France-Hawai‘i Telescope (CFHT) which is operated by the National ResearchCouncil of Canada, the Institut National des Sciences de l’Univers ofthe Centre National de la Recherche Scientifique of France, and theUniversity of Hawai‘i. (cid:63)(cid:63)
The full version of this paper with the appendices in available on line at ∼ dehrenre/articles/afsurvey/ sive enough for triggering the runaway accretion of hydrogen(see, e.g., Mordasini et al. 2009). On the other hand, ∼
3% ofplanets are detected by direct imaging searches at larger separa-tions ( > ∼ au ), preferentially around early-type stars. Most ofthese imaged planets seem to have a di ff erent origin, namely thefragmentation of the protoplanetary gaseous circumstellar disc(see, e.g., Dodson-Robinson et al. 2009; Kratter, Murray-Clay &Youdin 2010). Thus, there could be two possible paths of planetformation, depending on the stellar mass and planet separation.The core-accretion theory (Pollack et al. 1996) is currentlyprefered to the gravitational instability theory (Boss 1997) asthe origin mechanism of most planets detected by velocimetrybecause the latter mechanism is not expected to produce giantplanets at distances < ∼ au (Rafikov 2005). Meanwhile, obser-vational support has been gathered in favour of the core accretiontheory: the correlation between the planet frequency and the hoststar metallicity (Santos, Israelian & Mayor 2001, 2004) and theemerging fact that planets with an intermediate mass betweenNeptune’s and Saturn’s are exceedingly rare are both predictedin this theoretical frame. The recent extent of velocimetric sur-veys to low-mass main sequence stars (M dwarfs; Mayor et al.2009a,b) yielded new key evidence in favour of this formationmechanism: the fact, for instance, that low-mass planets (about a r X i v : . [ a s t r o - ph . S R ] J un Ehrenreich et al.: Deep imaging of close companions to A–F stars the mass of Neptune) are frequent while gas giants are seldomfound around low-mass stars observed by radial velocity or mi-crolensing techniques (Bonfils et al. 2007; Sumi et al. 2010). Onthe contrary, the frequency of giant planets should not be im-pacted by the stellar mass in the frame of gravitational instabil-ity, as long as circumstellar discs are massive enough to becomeunstable (Boss 2006).In fact, the total mass of the circumstellar disc where planetsform, is supposed to scale with the mass of the central star. Anincrease of the disc total mass can lead to an enhanced growthrate for protoplanetary cores in the disc midplane (Ida & Lin2004), leading to a larger number of massive planets aroundearly-type stars than around Sun-like stars. Hence, there shouldbe an observational correlation between planet occurrence andstellar mass (Laughlin, Bodenheimer & Adams 2004; Ida & Lin2005; Kennedy & Kenyon 2007). Surveying early-type, main-sequence stars hotter and more massive than the Sun should al-low to test such a correlation.Radial velocimetry can be used to detect planets at shortperiods around main-sequence early-type stars, as shown byLagrange et al. (2009a). These authors used the HARPS spec-trograph installed at the ESO 3.6-m telescope in La Silla (Chile)to survey a sample of 185 stars. They measured on each star theachievable detection limits, taking into account the ‘jitter’ levelof each object. These authors estimated that planets with periodsup to 100 days could be found around ∼
50% of the surveyedstars. Constraining the presence of planets at larger separationshowever requires a di ff erent detection technique.In this respect, direct imaging is a powerful tool used in com-plement to radial velocimetry for detecting companions at sep-arations typically > ∼ au . It is sensitive to intrinsically brightobjects: low-mass stars, brown dwarfs, and young giant plan-ets, and bring essential insights on the possible origins of theseobjects (Chauvin et al. 2010). Direct near-infrared imaging al-lowed observers to detect several kinds of substellar companionssuch as brown dwarfs (e.g., Nakajima et al. 1995; Lowrance etal. 1999, 2000; Chauvin et al. 2005a), objects at the transitionbetween brown dwarfs and planets (e.g., Chauvin et al. 2005b),or planetary-mass objects such as 2M1207b in orbit around abrown dwarf (Chauvin et al. 2004, 2005c); all of which are likelynot formed through core accretion like planets detected by ve-locimetry, but rather as multiple stars via a cloud fragmenta-tion process or as brown dwarfs through disc fragmentation (see,e.g., Lodato, Delgado-Donate & Clarke 2005). These detectionsnevertheless show that star-forming mechanisms can be e ffi cientdown to planetary masses.Recent breakthroughs in high-contrast imaging enriched thispicture, as giant planets were directly imaged around Fomalhaut(Kalas et al. 2008), HR 8799 (Marois et al. 2008), and β Pictoris(Lagrange et al. 2009b, 2009c, 2010); all these stars being youngand early main-sequence A stars. The discovery of a compan-ion to β Pic, a 9 ± J planet with a semi-major axis of ∼ au , implies that giant planets can indeed form in ∼
10 Myr. According to theoretical prediction of Kennedy &Kenyon (2007), this particular planet is close enough to its hoststar and could have formed in-situ by core accretion. This is notso clear for Fomalhaut b, which is located 115 au away fromits star, meaning that a particular migration scenario is requiredif the planet formed closer to the star via core accretion (Crida,Masset & Morbidelli 2009), or for the three planets or browndwarfs located at 68, 38, and 24 au from HR 8799. Gravitational The excess of scatter in radial velocity measurements resulting frominhomogeneities on the stellar surface and stellar envelope pulsations. instability, rather than core accretion, seems more suited toexplain the existence of such massive planets on wide orbits(Dodson-Robinson et al. 2009). The picture is however not thatsimple, as the gravitational instabilities leading to the fragmen-tation of massive circumstellar discs seem to produce objectswith masses typically above the deuterium-burning planetary-mass limit. In fact, according to Kratter, Murray-Clay & Youdin(2010), atypical disc conditions are required to form planetary-mass object via this mechanism. These authors suggest that ifthese planets formed this way, they must lie in the low-mass tailof the disc-born binary distribution. In this case, a larger numberof brown dwarfs or low-mass stars (such as M stars) should befound around A-type stars at distances of 50–150 au .In this frame, we perfomed a deep imaging survey of early-type A and F stars included in the velocimetric survey ofLagrange et al. (2009a). This study would allow us to testwhether low-mass stars and brown dwarfs commonly cohabitwith massive main-sequence stars, bringing new constraints onthe origin of the massive planets imaged around these stars. Inaddition, this imaging survey could help determining the impactof stellar multiplicity on the presence of closer-in planets, de-tected in parallel with radial velocimetry. Binarity is indeed an-other critical parameter for the theory of formation and evolutionof planets. In particular, the separation of the binaries could im-pact the way giant planets form, either by core accretion or discinstability (Zucker & Mazeh 2002; Eggenberger, Udry & Mayor2004; Desidera & Barbieri 2007; Duchˆene et al. 2010), and themodels of binary discs now predict observational e ff ects. Mayeret al. (2005), for instance, predict a lack of planets around eachcomponent of binary systems with separations < ∼ au becauseof instabilities in the circumstellar discs. The subsequent dynam-ical evolution also depends on the properties of the star and theouter companion (see, e.g., Rivera & Lissauer 2000).Only adaptive optics (AO) deep imaging allows to test prop-erly the presence of massive substellar companions. A num-ber of studies set to probe the existence and the impact ofsuch objects to exoplanetary systems detected by velocimetryhas been previously undertook (Patience et al. 2002; Luhman& Jayawardhana 2002; Chauvin et al. 2006, 2010; Mugrauer,Seifahrt & Neuh¨auser 2007; Eggenberger et al. 2007). For in-stance, Eggenberger et al. (2007) searched for bright long-periodcompanions around 130 G- and K-type stars and measured a bi-nary fraction of (8 . ± . (cid:48)(cid:48) (cid:48)(cid:48) au , whereas they found a higher binary fractionof (12 . ± . ζ Virginis (Hinkley et al. 2010 and referencestherein), in search of substellar companions.In this work, we are focusing on the close environment of asample of austral A- and F-type stars, which we mainly observedfrom the southern hemisphere using the AO system NaCo in-stalled at the ESO Very Large Telescope (VLT) on Cerro Paranal(Chile). Some austral stars were also observed from the north-ern hemisphere using PUEO, the AO bonnette of the Canada-France-Hawai‘i Telescope (CFHT) on Mauna Kea (USA). In afew cases, we also used NaCo to observe stars with positive de-clinations. This paper reports on these observations, which aredescribed in Sect. 2. The data reduction is detailed in Sect. 3, wepresent our analysis in Sect. 4 and discuss the results in Sect. 5. hrenreich et al.: Deep imaging of close companions to A–F stars 3
2. Star sample and observations
The present survey is the imaging part of the radial-velocity sur-vey described in Lagrange et al. (2009a) that was designed todetect planets around early-type stars with the HARPS spectro-graph at the ESO 3.6-m telescope in La Silla (Chile). The sur-vey is limited to dwarfs with spectral types ranging from F7 toB8 . Figure 1 presents the stars in a Hertzsprung-Russell dia-gram while Fig. 2 shows the number of stars observed per spec-tral type. In total, we surveyed 38 stars, including 16 F-, 19 A-,and 3 late B-type stars.The velocimetric and imaging surveys are also volume-limited, with distance limits set at 33 and 67 pc for the F0–F7and B8–A9 dwarfs, respectively. Figure 3 gives the number ofstars observed per 5-pc distance bin. The di ff erence in distancelimits would allow us to have roughly the same number of Aand F stars in the sample. Note, however, that one early F star(HD 4293) with a distance of 66.6 pc is included in the survey.Finally, the few stars present in our sample that are notincluded in Lagrange et al.’s (2009a) sample are part of thenorthern radial-velocity survey in progress with the SOPHIEspectrograph (Bouchy et al. 2009) at the 1.93-m telescope inObservatoire de Haute-Provence (France). The properties of allstars in the sample can be found in Table 1.While more details about the sample selection are given inLagrange et al. (2009a), we emphasize three biases. (i) First,spectroscopic binaries and close visual binaries with separationssmaller than 5 (cid:48)(cid:48) known at the beginning of the velocimetric sur-vey were excluded from the target list. Hence, these were notconsidered in the imaging survey as well. (ii) In addition, thepresent imaging survey is in fact strongly biased towards ‘in-teresting’ targets, i.e., stars around which companion candidates(CC) were detected during early epochs. This strategy was setmainly in order to compensate for the limited amount of observ-ing time devoted to the imaging programme. (iii) Finally, pooratmospheric conditions and technical problems that were experi-enced especially during observing runs VI and VII, prevented usfrom obtaining second-epoch observations for 13% of those tar-gets with companion candidate(s) identified during a first epoch;atmospheric conditions making the AO correction loop unstablealso prevented us from obtaining the highest achievable contrast– through the use of the coronagraph – for 10% of the targets. Data were recorded during seven observing runs (or epochs) per-formed between January 2005 and August 2009, at the VLT andCFHT, where we totalized 7 + We used the NAOS-CONICA instrument (NaCo) set on the VLTUnit Telescope 4 (Yepun) to benefit from both the high im-age quality provided by the Nasmyth Adaptive Optics System(NAOS; Rousset et al. 2003) at infrared wavelengths andthe good dynamics o ff ered by the Near-Infrared Imager and Velocimetric detection limits for stars earlier than B8 does not fallinto the planet domain.
Fig. 1.
Colours of stars in the imaging sample. Stars with spectral typesF, A, and B are represented red-, green-, and blue-filled circles, respec-tively.
Fig. 2.
Spectral types of the stars in the sample.
Spectrograph (CONICA; Lenzen et al. 2003) detector, in or-der to study the close circumstellar environment of 38 early-type stars. The NaCo Shack-Hartmann visible wavefront sensorwas chosen to perform the AO corrections on these bright tar-gets, used as self-references. NaCo allowed us to perform directimaging as well as coronagraphic imaging in order to improvethe image contrast. We used the coronagraphic mode consist-ing in a Lyot stop in the pupil plan of the telescope, combinedwith a occulting mask of diameter / (cid:13) = . (cid:48)(cid:48) λ c , bandpass ∆ λ , and transmission T ) are listed inTable 4. For instance, poor atmospheric conditions degrade theStrehl ratio S ; using a long-wavelength filter (typically K s ) al-lows to compensate for this degradation since the value of S also scales with wavelength as S = exp − (2 πω/λ ) , where ω isthe root-mean-square deviation of the wavefront.Two objectives were employed in order to optimize the pointspread function (PSF) sampling in these di ff erent bandpasses.The S13 camera has a field of view of 14 ×
14 arcsec and amean plate scale of 13.21 mas pixel − ; it was preferentially used Ehrenreich et al.: Deep imaging of close companions to A–F stars
Table 2.
Epoch dates and astrometric calibrations for all observing runs.Epoch UT Date Nights Instrument / Camera Filter Plate scale True north(mas) ( ◦ )I 2005-01-27 4 PUEO / KIR Br γ . ± . ± . / S27 J 27 . ± .
06 0 . ± . / S27 K s . ± .
06 0 . ± . / KIR Br γ . ± . − . ± . / KIR K 34 . ± . − . ± . / KIR Br γ . ± . − . ± . / KIR Fe ii . ± . − . ± .
02V 2008-08-20 2 NaCo / S27 K s . ± .
03 0 . ± . / S13 K s . ± . − . ± . / S13 H 13 . ± . − . ± . / S27 K s . ± . − . ± . / S13 H 13 . ± . − . ± . / S13 J 13 . ± . − . ± . / S27 K s . ± . − . ± . Fig. 3.
Number of stars per 5-pc distance bin. The F, A, and B stars arerepresented by the histograms filled with tight diagonal red stripes, di-agonal green stripes, and horizontal blue stripes, respectively. The dot-ted line stands for the whole star sample. when observing in J and H bands. When observing in the K s band, the S27 camera was chosen: it o ff ers a 28 ×
28 arcsec fieldof view and a mean plate scale of 27.06 mas pixel − . For eachepoch, precise plate scales were redetermined using astrometriccalibrators (see Table 2).Our observing strategy consists, for each target, in obtain-ing a first-epoch image and, when companion candidates are de-tected, to perform a second-epoch observation in order to testtheir status: either field objects (background contaminants) orcomoving objects (physically bound to the targeted star). Thediscrimination between these two possibilities is allowed by thestellar proper and parallactic motions in the plane of the sky. Thisis illustrated in Fig. 4. Hence, depending on motion amplitudes,a few months to a few years are necessary between the first- andsecond-epoch observations. For instance, for two observationsperformed three years apart, and a precision on the measured po-sitions of individual point sources on each epoch of ∼ σ discrimina-tion between comoving and non-comoving sources is possiblefor typical minimum proper motions of ∼ − .A target is first observed both in direct imaging and corona-graphic modes. For epochs V, VI, and VII, the CoroObsAstro observing template was used: this template consists in acquiringfirst some coronagraphic exposures on the object without anyneutral density filter (
Full Uszd setting, where the pupil area issimply reduced by the presence of the Lyot stop, which blocks14% of the light). Then, a neutral density filter (
ND Short ) is in-serted in the focal plane while removing both the coronagraphicmask and the Lyot stop in the focal and pupil plan, respectively.A direct image is then taken, the object being at the same exactlocation as behind the mask. This allows for retrieving the starposition behind the mask, enabling precise astrometric measure-ments of separation with companion candidates only visible incoronagraphic mode. The star is then jittered (by steps of ∼ (cid:48)(cid:48) )on 4 di ff erent locations on the detector in order to obtain thebackground sky estimation for the direct images. The last o ff setsends the telescope ∼ (cid:48)(cid:48) away from the target star, and the oc-culting mask is used again – the neutral density filter is removedonce more – as 5 jittered exposures are taken on the sky in orderto estimate the sky background for coronagraphic images. Thistemplate was not available during our first NaCo observing run(epoch II), hence the precise location of the star behind the maskis not known, and the astrometry is therefore less precise thanfor later observations. In all cases, the same amount of time wasspent on the object and on the sky.A systematic astrometric shift is induced when switching theneutral density filter back and forth: we used pinhole calibrationsrecorded at each epoch to ensure that this shift is small: typically ± . ± .
01 pixel along the detector x - and y -axes, re-spectively. This shift is included in our error budget within the σ s term of Eq. (3).When a companion candidate can be detected without thecoronagraph, the star is reobserved at a later epoch, only in directimaging mode. NaCo observing templates ImgGenericOffset and
ImgAutoJitter were used for this purpose.
PUEO (Rigaut et al. 1998) is the AO bonnette mounted on the3.6-m CFHT. It was used in combination with the near-infraredcamera KIR (Doyon et al. 1998) which has a field of view of35 ×
35 arcsec , and a mean plate scale of 34 . − .Depending on atmospheric conditions and source brightnesses,we used broad- and narrow-band filters listed in Table 4.We performed unsaturated direct imaging to investigate asclose as possible to the star, while the detector was saturated inorder to increase the contrast at larger separations (there is no oc- hrenreich et al.: Deep imaging of close companions to A–F stars 5 Table 4.
Properties of the filters used during the survey.Filter λ c ( µ m) ∆ λ ( µ m) T (%)VLT / NaCoJ 1.27 0.25 78H 1.66 0.33 77NB . . s / PUEOH 1.632 0.296 85Fe ii (v = γ culting mask on PUEO). A special care was taken to ensure thatthe unsaturated exposures could be used as references to mea-sure the relative position and photometry between the star andthe companion candidate(s). Images are jittered on the detec-tor so that the sky background can be retrieved. The observingmethod is otherwise similar to that described above for NaCo.The 6 stars from our sample with PUEO epochs have all beenobserved with NaCo, and are thus included in this study for con-sistency.
3. Data reduction
Flat fields have been taken on the twilight sky for each observ-ing run at VLT (usually within a few days from the observingnights). At CFHT, dome flat fields were taken for every filter atleast once per run.A well-known field in the Orion Trapezium centered onthe star θ Orionis C was used for astrometric calibrations(McCaughrean & Stau ff er 1994). One observation per fil-ter / camera set-up used was taken at each observing epoch todetermine the mean plate scale and the true north orientation.On each observation, the exposure time was adjusted to satu-rate the 2 or 3 brightest stars of the field so that from 15 to 40stars (depending on the size of the field, from 13 (cid:48)(cid:48) with the S13camera of NaCo to 35 (cid:48)(cid:48) with PUEO) have a su ffi cient signal-to-noise ratio ( ≥ s x , s y ) of eachnon-saturated star on the reduced image of the field is measuredin pixels; these values are then compared to the position on thesky ( ρ x , ρ y ) in arcseconds reported by McCaughrean & Stau ff er(1994) with the following equations: ρ x = x + p x s x cos θ n + p y s y sin θ n (1) ρ y = y − p x s x sin θ n + p y s y sin θ n (2)where p x ( p y ) is the plate scale along the x - ( y -) axis, θ n is theorientation of the detector on sky, and x and y are o ff sets giv-ing the correspondance between the absolute positions (it is onlyused to solve the equation). A Levenberg-Marquardt minimiza-tion is then applied to find the plate scale solution. The errors onthe plate scale and orientation are given by the standard devia-tion of the corresponding parameter. The mean plate scales andorientations are reported in Table 2. The reduction procedure makes use of the eclipse utility li-brary (ESO 1996–2002; Devillard 1997). For each observingtemplate on a target, coronagraphic exposures are sorted by eye.Good exposures (i.e., exposures where the star is well occultedand the AO correction loop is closed) are then concatenated intoa data cube. A coronagraphic sky is created as the median of alldithered coronagraphic exposures on the sky, and is correctedtwice for bad pixels detected (i) on the reduced flat field and (ii)on the median sky once the first bad-pixel rejection as been ap-plied. The cleaned median sky is then subtracted to each planeof the data cube, which are next divided by the flat field. The 2bad-pixel maps are again used on the flat-fielded data cube. Thefinal reduced image is extracted as the median of the data cube.
The idea is similar to the reduction of data taken in corona-graphic mode, except that all the steps are performed by the jitter routine, which realigns dithered direct images via across-correlation algorithm and does all the cosmetic tasks. Anadditional (but similar) reduction is devoted to the first direct im-age taken after the coronagraph is removed (see Sect. 2.2.1); thisreduced image will be used as an astrometric reference for theposition of the star behind the mask.
4. Analysis
Candidate companions are spotted by a close eye-examination ofeach star, and their absolute position on the detector determinedto a ∼ IDL wrapper pro-gramme, that successively performs three astrometric and pho-tometric estimations. An aperture estimation (i), where the aper-ture is centered on the input position, returns the centroid po-sition as the barycentre of the light within the aperture, and anarbitrary magnitude estimation from the integration of all pix-els within the aperture, weighted by the flux of the referencePSF within the same aperture. The position ( ∆ x , ∆ y ) and magni-tude ∆ m of each companion candidate are then given relativelyto those of the star. A Gaussian fitting (ii) using the IDL rou-tine
GCNTRD is also applied to each input position and returnsrelative separations. Finally, (iii) a deconvolution algorithm us-ing a Levenberg-Marquardt minimization of the log-likelyhoodfunction (V´eran 1997) is employed: it basically consists in a PSFfitting within the pupil plane. Procedures (i) and (ii) work wellwhen the objects are well separated on the image, and give sim-ilar results than procedure (iii). However, they do not give accu-rate results for tight systems (typically tighter than the apertureradius used). In these cases, our estimations rely on procedure(iii).The astrometric precision is ultimately limited by the astro-metric calibrator, and the precision σ p (in mas) obtained on theestimation of the plate scale p (in mas pixel − ) at each epoch(which are listed in Table 2). Yet, our procedure for extractingthe astrometric parameters of an image is also characterized byan uncertainty σ s on the measured separation s (both in pixels). Ehrenreich et al.: Deep imaging of close companions to A–F stars
The total uncertainty on the measured separation ρ (in mas) isthus σ ρ ρ = (cid:115)(cid:18) σ s s (cid:19) + (cid:32) σ p p (cid:33) . (3)The values of σ s depend on various parameters such as the actualseparation of the components as well as their contrast. In the fol-lowing, we choose to take σ s = . ffi cultcases, such as the very tight companion to HD 43940.Since the stellar flux is always estimated from the direct (un-saturated) image, both the direct and coronagraphic images arenormalized by their respective exposure times when the pho-tometry is performed for objects present in NaCo coronagraphicimages. The magnitude di ff erence is then increased by 4.6 magto take the transmission di ff erence introduced by the use of aneutral density filter into account. In the case of NaCo corona-graphic imaging, the use of a slightly undersized aperture due tothe presence of the Lyot stop in the pupil plan leads to an ad-ditional uncertainty of ∼
4% on the fluxes measured when theocculting mask is on, with respect to the full aperture availablewhen it is o ff (Boccaletti et al. 2007).As for the separation, the determination of the contrast be-tween two components is also impacted by their respectivebrightnesses and separation. We can actually estimate our pho-tometric precision from the set of dispersion of the measure-ments taken on the same objects at di ff erent epochs: the mediandispersion obtained is 0.4 mag (45% on the measured fluxes).Astrometric and photometric measurements for all companioncandidates imaged are given in Table 5. Stars without any de-tected companion candidates are listed in Table 6. The companionship, i.e., the fact that a candidate companionis comoving with the primary star in the vincinity of which ithas been imaged – which is indicative of a physical associa-tion bound by gravity – is appreciated by measuring the evo-lution of the separation ρ = √ ( ∆ x + ∆ y ) and the position angle θ = arctan ∆ x / ∆ y (from north to east) between the candidate andthe star over several epochs of observation. The values of ∆ x and ∆ y are corrected for the orientation θ n of the detector north withrespect to the true north (given in Table 2 at each epoch). Theshift in the position (right ascension α and declination δ ) of acompanion candidate between two epochs i and j is ∆ α i → j = ρ j sin θ j − ρ i sin θ i (4) ∆ δ i → j = ρ j cos θ j − ρ i cos θ i . (5)These measured shifts are compared to the theoretical appar-ent motion the object should have on the sky, given the stellarproper motions listed in Table 1, the parallactic motion, and as-suming, as standardly done, that the candidate is a motionlessbackground contaminent. Figure 4 shows two examples of com-panionship determination. The positions of all companion can-didates obtained during several observing runs are representedin Fig. A1 (see the Appendices ). The full version of this paper with the appendices in available on line at ∼ dehrenre/articles/afsurvey/ For observations performed at N epochs, it is possible to cal-culate the probability P bkg that a companion candidate is a back-ground object, using a χ probability test of 2 N − χ = N (cid:88) i = (cid:16) ∆ α → i − ∆ α (cid:63) → i (cid:17) σ ∆ α → i + σ ∆ α (cid:63) → i + (cid:16) ∆ δ → i − ∆ δ (cid:63) → i (cid:17) σ ∆ δ → i + σ ∆ δ (cid:63) → i , (6)where ∆ α (cid:63) → i and ∆ δ (cid:63) → i are the theoretical shifts in α and δ inthe position of the star between epochs 1 and i , taking the stellarproper motion and the parallactic motion into account.The probability of observing a value of χ that is larger thanthat obtained with Eq. (6) for a random sample of N observa-tions with ν = N − χ -distribution (Bevington & Robinson2003), P χ ( χ ; ν ) = ν/ Γ ( ν/ (cid:90) + ∞ χ (cid:16) x (cid:17) ( ν − / e − ( x ) / d (cid:16) x (cid:17) (7)where the Γ ( n ) function is equivalent to the factorial function n ! extended to nonintegral arguments, and ( x ) represents thepossible values of χ within the integral sum. The probability P bkg is obtained from Eq. (7); practically, we are using the IDL chisqr pdf function. A status of background contaminant isassigned to each object for which P bkg > . ff ects in the position measurements,due for instance to variations in the stellar proper motions, dis-crepancies between the real values and those from the literature,or the fact that background objects have a non-negligible propermotion on the sky, we calculated P bkg < .
01% for some objectsthat are evidently not comoving with the primary star. This is,for instance, the case of HD 68456 CC P cmv , also given in Fig. A1, that a companion candidate iscomoving with the star, using a test similar to Eq. (6), χ = N (cid:88) i = (cid:32) ∆ α → i σ ∆ α → i (cid:33) + (cid:32) ∆ δ → i σ ∆ δ → i (cid:33) . (8)The result is injected into Eq. (7). Note, however, that the valueof P cmv can only be used to give a hint of a physical associa-tion . In fact, this probability does not take into account the or-bital motion of a true comoving companion. Hence, an objectwith very small values of P bkg and P cmv ( < . P bkg < .
01% and P cmv > .
01% is comovingwith the star: 4 companion candidates out of 41 detected pointsources (10%) are confirmed comoving objects according to thistest (they are labelled ‘(C)omoving’ in Table 5). Consequently,an object with P bkg < .
01% and P cmv < .
01% could be con-sidered to be a background contaminant. On this basis, we canconfidently reject 6 additional point sources (15%).Another tricky case is that of HD 41742 CC P bkg = P cmv = hrenreich et al.: Deep imaging of close companions to A–F stars 7 Table 6.
Sample stars around which no companion candidates were detected.*At CFHT, saturated imaging was used instead of coronagraphic imaging. × NDIT × NEXPDirect imaging Coronagraphic imaging1 HD 2834 2005-11-06 K s S27 0 . × ×
10 3 . × ×
222 HD 3003 2005-11-07 K s S27 0 . × ×
10 6 . × ×
303 HD 4293 2005-11-08 K s S27 1 . × ×
10 10 . × ×
204 HD 10939 2005-11-06 K s S27 0 . × ×
10 5 . × ×
207 HD 16754 2005-11-07 K s S27 0 . × ×
10 5 . × ×
208 HD 19545 2008-08-20 K s S27 2 . × × s S27 0 . × ×
10 2 . × × s S27 0 . × ×
10 8 . × × . × ×
18 30 . × × s S27 3 . × × s S27 1 . × × s S27 1 . × ×
10 5 . × × s S27 1 . × × s S27 0 . × ×
10 1 . × × s S27 0 . × × s S27 0 . × ×
10 1 . × × s S27 0 . × × . S27 2 . × ×
10 1 . × × s S27 5 . × × well compatible with a comoving object (see HD 41742 CC For each image, we derive 6- σ detection limits in the form of a2-dimensional map. The detection limits are calculated by mea-suring for each pixel the noise (standard deviation) in a 5 × IDL routine
IMAGE VARIANCE (by M. Downing). The coronagraphic images are normalized tothe exposure time of the associated direct image before theirdetection limits are calculated; the neutral density filter trans-mission is also taken into account. The classic 1-dimensionaldetection limits are then radially extracted from the 2D maps:the 1D detection limit at a separation ρ from the star positionat ( ∆ α, ∆ δ ) = (0 ,
0) is the azimuthal median of all pixels in anannulus of mean radius ρ with 5-pixel thickness.These 1D detection limits can be used for determining theoverall properties of our survey and comparing it to other sur-veys. However, since the contrast is mainly limited by the pres-ence of speckles at close separations, we believe that the 2Dmaps are more accurate than the 1D limits and should rather beused when refering to particular objects in the survey. At largeseparations from the central stars, the contrast is rather readout-noise limited, and both 1D or 2D limits could be used.Figures 5 and 6 show the 2D- and 1D-detection limitsachieved both in direct imaging and coronagraphic modes for two stars from our survey: an early A-type star (HD 2834; Fig. 5)and a late F-type star (HD 91889; Fig. 6) are chosen as typi-cal examples. Maps of the detection limits for all surveyed starsare available in Appendix B. The overall 1D detection limits inthe K s band are presented in Fig. 7. Unlike the direct imag-ing observations, the coronagraphic imaging mode did not usea neutral density filter, and therefore enabled improved sensi-tivities for wider-separation (greater than 2 (cid:48)(cid:48) ) sources. With thecoronagraph in place, we were able to employ large DetectorIntegration Times (DITs), while still avoiding detector satura-tion. The larger DITs helped reduce detector read-out noise,which currently limits the wide-separation sensitivities. At sep-arations greater than ∼ (cid:48)(cid:48) , we measured ∼ ff erent spectral types of stars included in the sur-vey have (i) di ff erent distance cut-o ff values (see Fig. 3) and (ii)di ff erent median ages, it is interesting to plot the overall detec-tion limit for each spectral type (A, F, and B) as the median ofall detection limits. Note that since there are only 3 B stars inthe survey, the median limit for B stars (blue curves) cannot beconsidered as representative as for A and F stars (red and greencurves, respectively).In this scope, the performance in contrast is slightly better,typically by < ∼ . . (cid:48)(cid:48)
2, 0 . (cid:48)(cid:48)
5, and > (cid:48)(cid:48) from the star, respectively.The use of an occulting mask allows to obtain better contrastthan with direct imaging from ∼ . (cid:48)(cid:48)
5. A maximum contrast of ∼
13 mag is reached at separations larger than 5 (cid:48)(cid:48) from the star.
Companion candidates are detected at all separations above thedetection limits. Table 5 presents the observing dates, integrationtimes, projected separations, position angles, and magnitude dif-ferences of the 41 point sources we identified throughout our ob-serving campaign. Based on the method presented in Sect. 4.2,we were able to determine that 20 objects (49% of sources de-tected) were background sources and 6 objects (15% of sources
Ehrenreich et al.: Deep imaging of close companions to A–F stars
Fig. 7.
Median 6- σ detection limits in the K s band for F- (red), A-(green), and B-type (blue) stars, in direct imaging (plain lines) and coro-nagraphic (dashed lines) modes. The coronagraphic detection limits arebetter than the direct imaging ones from 2 (cid:48)(cid:48) on, for the reasons detailedin Sect. 4.3. The measured separations ρ and magnitude di ff erences ∆ K s of companion candidates confirmed through multi-epoch observationsand those with an undefined status are reported by filled and empty cir-cles, respectively. The confirmed companion candidate to ι Crt, whichwas only observed in the H band, is marked by the red square assuminga ( H − K ) colour close to 0. The dash-dotted line at ρ = . (cid:48)(cid:48)
35 indicatesthe occulting mask radius.
Fig. 8.
Absolute magnitudes vs. projected separations. The legend isthe same than for Fig. 7. The absolute magnitudes M K s and projectedseparations are calculated for each object given its apparent magnitude m K s in the K s band and distance d from Earth. The 6- σ detection limitsshown for F- (red), A- (green), and B-type (blue) stars, in direct imaging(plain lines) and coronagraphic (dashed lines) modes are the medians ofthe detection limits M K s = f ( ρ ( d ) , d , m K s ) calculated for all stars of agiven spectral type. detected) were in fact comoving with their primary stars. Theseconfirmed companions are shown in Fig. 9 and their physicalproperties summarized in Table 7. Because a second epoch ob-servation could not have been obtained for all targets, 15 pointsources are still of undefined nature. In the following, we detailthe properties of all confirmed companions as well as those of a selection of remarkable cases of companions with undefiniedstatus and some that we rejected as comoving objects. For companions comoving with their primary stars, the projectedphysical separation in astronomical units can be simply extractedfrom the knowledge of the separation measured on the sky andthe parallax (usually, the
Hipparcos parallax; ESA 1997). Themasses of companions can be retrieved using mass-magnituderelationships yielding from evolutionary models. The detectionlimits of our survey of A and F stars (Fig. 7) show that weare sensitive to companions in the brown-dwarf to low-massstar regimes. Thus, we chose the evolutionary models for solarmetallicity low-mass stars and the associated mass-magnituderelationships provided by Bara ff e et al. (1998) to determine themasses of our detected comoving companions. Pratically, themasses were obtained for given absolute magnitudes and agesby interpolations into Bara ff e et al.’s tables (VizieR On-line DataCatalogueJ / A + A / / HD 14943—
A companion with an absolute magnitude of M K = . ff erent epochs2 . (cid:48)(cid:48) au ) away from the A5V star HD 14943 (HIP 11102).According to Bara ff e et al.’s (1998) evolutionary models forsolar-metallicity low-mass stars, this companion should be a0.4 M (cid:12) star, assuming an age of 850 Myr, determined in theframe of Su et al.’s (2006) survey for debris discs around A starswith Spitzer / MIPS. Hence, HD 14943B, shown in Fig. 9a, couldbe a main sequence M2 star (Cox 2000). HD 41742—
This high proper-motion star is part of a physicalternary hierarchical system (Multiple Star Catalogue — MSC—, Vizier Online Data Catalogue J / A + AS / /
75, Tokovinin1997). We identify CC1 to component B (CCDM J06046-4504B; see Fig. 9b), a 0.79 M (cid:12) star with a period of ∼ B − V = .
01, typical of a K0 / K2 star and according to Cox(2000). The two other companion candidates (CC2 and CC3) arebackground objects. The component C of the system cataloguedin the MSC is too far to enter NaCo or PUEO field-of-view( ρ C = . (cid:48)(cid:48) . (cid:48)(cid:48)
9, i.e., 159 au and an abso-lute magnitude M K = . (cid:12) star using Bara ff e et al.’s (1998) mass-magnitude relationships,in good agreement with the mass previously determined. This es-timation was made assuming an age of 3.7 Gyr, according to theGeneva-Copenhagen Survey of Solar Neighbourhood (GCSSN,Vizier Online Data Catalogue V / HD 49095—
This high proper-motion star with radial veloc-ity variations (Lagrange et al. 2009) is referenced as a binary byMakarov & Kaplan (2005) based on the di ff erence of proper mo- According to Drilling & Landolt’s calibration of MK spectral types(in Cox 2000, p. 389, Table 15.8). This star exhibits radial velocity vari-ations (Lagrange et al. 2009).hrenreich et al.: Deep imaging of close companions to A–F stars 9 Fig. 10.
The F6.5V star HD 49095 and its binary comoving companioncomposed of CC1 (blue circle) and CC2 (red circle). The axis are la-belled in projected separation with respect to the primary star centroidposition (white star). The positions of objects on this image are retrievedusing the deconvolution algorithm described in Sect. 4.1. tions measured between two di ff erent catalogues (‘ ∆ µ binary’),namely, the Hipparcos and Tycho-2 catalogues. Here, we im-age the binary component and resolve it as a tight binary inwhich orbital motion is detected over several epochs. The sys-tem is shown as seen on 2005-11-07 in the K s band with NaCoin Figs. 9c and 10 and the revolution of the B (CC1) and C (CC2)components around their center of mass is shown over 3.5 years(4 epochs) in Fig. C1. The position of HD 49095C with respectto HD 49095B was retrieved using the deconvolution algorithmdescribed in Sect. 4.1 and is plotted in Fig. 11. For the twoVLT images (taken on 2005-11-07 and 2009-04-26) where theHD 49095BC system is well resolved, we assumed 1- σ positionerrors of 0.25 pixel. A position error of 0.5 pixel has been as-sumed for CFHT images (taken on 2007-01-27 and 2007-11-26)where the system is barely resolved. The tight system projectedseparation is seen to vary between 1.4 and 2.6 au . Yet, becausewe do not have enough points to fit a precise orbit, we simplycalculated a best-fit circular orbit, assuming the system is seenface-on, and determined a semi-major axis of 2.27 au ( χ = . (cid:12) . This is about threetimes smaller than the total mass determined photometrically us-ing Bara ff e et al.’s (1998) models (0 . + . = .
305 M (cid:12) )given the measured magnitudes in the H band of M HB = . M HC = . HD 101198—
The ι Crateris (HD 101198; CCDM J11387-1312A) entry in the CCDM mentions a companion (KUI 58B) at
Fig. 11.
Motion of the binary companion CC2 of HD 49095 (zoom fromFig. 10) relatively to CC1 (blue point) between 4 di ff erent epochs cov-ering 3.5 yr. The best-fit circular and face-on orbit is figured, with asemi-major axis of 2.27 au . ρ = . (cid:48)(cid:48) θ = − ◦ in 1934, with a visual magnitude V = ∆ µ binary. Were KUI 58B abackground contaminant, it would lie in 2009 at ρ ≈ . (cid:48)(cid:48)
45 and θ ≈ −
141 deg, so that we should be able to detect it on NaCo K s images, unless its V − K colour index is below ∼ − .
5. The otherpossibility is that KUI 58B is indeed bound to ι Crt. In this case,it is likely that CC1, shown in Fig. 9d, is that object. This wouldimply for CC1 colour indexes V − K ≈ − . = .
75 and V − H ≈ − . = .
56, typical of a M3 dwarf (Ducati et al.2001) . The mass derived from the measured M H = . (cid:12) , more typical of a M0 or a late K star (Cox2000). Furthermore, the primary shows radial velocity variations(Lagrange et al. 2009). HD 153363—
26 Ophiuchi (HD 153363) is referenced asa radial-velocity variable (Lagrange et al. 2009) as well asan astrometric binary (Proper motion derivatives of bina-ries, Vizier Online Data Catalogue J / AJ / / / A + A / / Here, we assume that the binary has the same age as the primary( ∼ Table 7.
Physical properties of comoving companions. au ) (M (cid:12) )5 HD 14943 1 141 0.4015 HD 41742 1 159 0.8017 HD 49095 1 32.3 0.17517 HD 49095 2 31.9 0.1322 HD 101198 1 25.0 0.5725 HD 153363 1 11.3 0.7036 HD 220729 1 14.9 0.30 It can be seen in Fig. A1 that the position of CC1 in April 2009,imaged in Fig. 9e, is not compatible with that of a backgroundfield object. Hence, the motion of CC1 can rather be explainedby the orbital motion around 26 Oph. We observed 26 Oph Bin three band passes and measured ∆ K s = . ∆ H = .
7, and ∆ J = .
1. Given the 2MASS
JHK magnitudes of 26 Oph andits distance of 33.3 pc, the companion has absolute magnitudes M K s = . M H = .
9, and M J = .
5. These magnitudes arecompatible, within the given uncertainty range of ± . (cid:12) object according to Bara ff e etal.’s (1998) model. Assuming the J − K s and H − K s coloursare each overestimated by ∼ . σ variation), they in-deed correspond to the colours of an early K dwarf (Ducati etal. 2001). Additional fainter and farther components of this sys-tem may exist; however, the confirmation of their companion-ship would require a second epoch with coronagraphic observa-tion. Meanwhile, we can already exclude CC2, also seen on thedirect images, as an additional physical companion to 26 Oph. HD 220729—
Lagrange et al. (2009a) identified o Gruis(HD 220729) as a spectroscopic binary and noted that it wasassociated to a ROSAT source by Suchkov, Makarov & Voges(2003). In this work, we confirm the binary nature of this staras we resolve it as a tight (0 . (cid:48)(cid:48) au ) and contrasted( ∆ K = .
7) binary system, as can be seen in Fig. 9f. We esti-mate the absolute magnitude of o Gru B to M K = . (cid:12) , i.e.,a M3 / M4 main sequence star. The companion detected in theadaptive-optics survey is not responsible for the radial-velocityshifts, hence o Gru is a triple system.
HD 32743— η Pictoris (HD 32743, HIP 23482) is cataloguedin the CCDM as a binary with an astrometry recorded in 1946of ρ = . (cid:48)(cid:48) θ = − ◦ , and a photometry of ∆ V = . ρ = . (cid:48)(cid:48) ± . (cid:48)(cid:48) . θ = − . ± . ◦ . We mea-sured infrared apparent magnitudes of J = . ∆ J = . K s = . ∆ K s = .
4) for CC1, which seem compati-ble with the visible apparent magnitude ( V =
13) of the knowncompanion to η Pic. Hence, it is likely that CC1 is the binarycompanion to η Pic and catalogued as CD-49 1541B. To es-timate whether CC1 is comoving with η Pic, we produced aproper motion plot similar to those of Fig. A1 and that is pre-sented in Fig. D1. We assumed arbitrarily large error bars onthe 1946 astrometric parameters tabulated in the CCDM. The position of the object measured in 2005 corresponds to that ofa background object, non-comoving with the primary star. Weobtain P bkg = . / CD-49 1541B; we consider it as a probable back-ground contaminant. We are nevertheless resolving it as a tightapparent binary and, since a second tight binary system is alsopresent in the surroundings of η Pic, the system will need to bemonitored in the future.
HD 43940—
Clearly seen in 2005-November observations inthe K s band, the close companion candidate to HD 43940 ( ρ = . (cid:48)(cid:48)
22) is harder to resolve from the primary star in 2008-August,both in K s and H bands. This may be due, however, to thepoorer observing conditions compared to 2005-November. A fi-nal attempt to characterize this new tight system was done on2009-April, but the companion could barely be resolved fromthe primary star in the J band. HD 43940 has very low propermotions in the direction of CC1 ( δ α = − .
51 mas yr − , δ δ = − .
63 mas yr − ), but since the separation is small, it is hard todisentangle between the two possibilities: (i) the system is ap-parent, CC1 is a nearly motionless background contaminant ap-parently getting closer to HD 43940 because of this star propermotions, or (ii) the system is physically bound and the di ffi cultyto resolve it as time passes results from the orbital motion ofCC1 around the primary. HD 158094—
The B8 star δ Aræ (HD 158094) is part of an op-tical (Torres 1986) double system (CCDM 17311-6041) whosesecondary component lies out the field of view ( ρ = . (cid:48)(cid:48) θ = ◦ ). Concerning the star itself, Spitzer / MIPS 24- µ m pho-tometry indicates an age of 125 Myr and nonexistent infrared ex-cess (Rieke et al. 2005). Low-mass companions were searchedfor around this star by Hubrig et al. (2001) using the AO sys-tem ADONIS on the ESO 3.6-m telescope in La Silla. Theseauthors report detection limits of ∆ K ∼ ∆ K ∼ ∼ (cid:48)(cid:48) –5 (cid:48)(cid:48) and > ∼ (cid:48)(cid:48) , respectively. This meansCC3 ( ρ = . (cid:48)(cid:48) θ = . ◦ , ∆ K = .
2) could have been de-tected by Hubrig et al. (2001): it is slightly above their detectionlimit, and close to the edge of, yet within, their 24 (cid:48)(cid:48) × (cid:48)(cid:48) field ofview, centered on the star, in 1999-March. Note that if CC3 werea background contaminant, then given the star proper motion, itwould have been closer to star in 1999 ( ρ = . (cid:48)(cid:48) HD 3003— β Tucanæ (HD 3003) was previously known asa multiple star (CCDM J00327-6302AB). Were this system co-moving, we should be able to retrieve both components, sepa-rated by 6 . (cid:48)(cid:48) HD 50445—
The star HD 50445 is not catalogued as being partof a multiple apparent or physical system. However, it lies at arelatively low Galactic declination ( b = − . ◦ ) and at least 5point sources have been detected in 2005-November in the 27 (cid:48)(cid:48) -wide field of NaCo S27 camera, using the coronagraphic mask.All 5 objects were observed again in 2009-April; a software is-sue during the observations unfortunately prevented recordinga direct image of the primary star at this epoch, hence the as- hrenreich et al.: Deep imaging of close companions to A–F stars 11 trometric precision is lower for the 2009 observations (typically1 pixel, i.e., 30 mas). We are nevertheless confident that all can-didate companions are background contaminants. Note that CC2can be seen in the infrared Digitalized Sky Survey 2 image takenon 1985-December at ρ ∼ (cid:48)(cid:48) and θ ∼ ◦ (the apparent mo-tion of the background contaminants on the sky between 1985and 2005 is µ α ∼ + . (cid:48)(cid:48) µ δ ∼ + . (cid:48)(cid:48) HD 177756— λ Aquilæ (HD 177756) is one of the B stars in-cluded in the Hubrig et al. (2001) survey. These authors did notreport any companions in the K band with a magnitude di ff er-ence up to ∆ K ∼
9. The observed CC1 has a magnitude di ff er-ence of ∆ K =
4. It should have been seen in the 1999 ADONISimage, as this background contaminant was closer to the star atthis epoch, unless the field of view of 24 (cid:48)(cid:48) × (cid:48)(cid:48) reported byHubrig et al. is not squarred (depending on the dithering patternused by these authors). HD 224392—
Members of the Tucana association, includingthe Vega-like star η Tucanæ (HD 224392; see Mannings &Barlow 1998 for the infrared excess), were surveyed in the in-frared by Neuh¨auser et al. (2003) with the SHARP-I infraredimager on the New Technology Telescope (NTT) at La Silla.These authors report detection limits in the K band of ∆ K = . > au from the star. However, they could not have de-tected CC1 for it was out of SHARP-I 13 (cid:48)(cid:48) × (cid:48)(cid:48) field-of-view.Zuckerman, Song & Webb (2001) note that as part of the IRASFaint Source Catalogue (Vizier online data catalogue II / η Tuc might also be a multiple star, yet in thiscase we do not resolve it. Besides, being a HARPS constant(Lagrange et al. 2009a) makes it unlikely to be a spectroscopicbinary. We do detect two companion candidates far from the star,which are not comoving with it.
5. Discussion
We found that 6 stars out of 38 in the sample have confirmedlow-mass companions, yielding an observed multiplicity frac-tion of 16% for our sample. This number can be seen as a lowerlimit value for the sample (i) because we have imaged compan-ion candidates around 7 other stars in the sample, without havingthe possibility to determine whether they are true companionsor background objects, (ii) because we could have missed closebinaries below our detection limits, and / or (iii) because somewide-separation binary components could have been too closeto the line of sight to be resolved at the time of observations. It isalso certainly a lower-limit value regarding the global multiplic-ity fraction of early-type stars, since our survey is biased againstspectral binaries or visual binaries with companions at ≤ (cid:48)(cid:48) , asdescribed in Sect. 2.1.In the case all the undefined candidates are bound to theirprimary stars, the multiplicity fraction of sample stars could beas high as 34%. More specifically, the multiplicity fraction forF stars in the sample is 5 / ≈ / ≈
5% forA + B stars. If we assume that unconfirmed companions are allphysically bound companions, then these fractions would riseto 50% and 23% for F and A + B stars, respectively. The multi-plicity fraction could be even higher than these figures due tonarrow-separation binaries and orbital projection e ff ects.Considering confirmed companions only, it would seem sur-prising that the multiplicity fraction of F stars is 6 times larger Fig. 12.
Number of detected companions as a function of the projectedseparation to the primary, per bin of 100 au . Companions around F stars(red stripes) and A + B stars (green stripes) are di ff erenciated. The blackdotted line represents the total number of companions. (a) Confirmedcompanions only. (b) Confirmed and unconfirmed companions. than that of A stars, whereas it is usually expected that the mul-tiplicity fraction increase with the stellar mass. Actually, this is abias that can be attributed to the fact that we are probing di ff erentphysical separations for di ff erent stellar types. In order to haveabout the same number of F and A stars included in this volume-limited survey, we have to go farther to include as many A starsas F stars: the median distance of sample F stars is 27.0 pc whileit is 54.9 pc for sample A and B stars (see Fig. 3). Thus, we areprobing smaller physical separations for F stars than for A andB stars, as shown in Fig. 7b. Five out of 6 confirmed compan-ions detected next to F stars are located between 10 and 40 au ,at projected separations not or hardly reached for the A stars inthe sample. On the contrary, a confirmed companion located at > au was found for each type of stars.Figure 12 shows the number of confirmed companions(Fig. 12a) or confirmed and unconfirmed companions (Fig. 12b)per bin of 100 au from their primary stars. The bias is clearlyapparent in Fig. 12a if we assume that our detection e ffi ciencyat <
100 AU is lower for A stars than for F stars. If we furtherassume that all unconfirmed companions are physically boundto their primary stars, then Fig. 12b shows that we have roughlythe same number of objects between 100 and 500 au , while com-panion candidates at projected distances greater than 500 au areonly found around A + B stars. It is likely that at least some un-confirmed objects detected around A stars are real companions,and contribute with the observational bias to fill the gap betweenthe multiplicity fractions of A and F stars in the sample.
The objects from the imaging survey presented in this work havebeen – or are being – monitored spectroscopically, seeking forradial velocity variations that would reveal the presence of close-in companions (see Lagrange et al. 2009). Two questions arisewhen companions are imaged around stars included in the ve-locimetric survey.First, is the spectroscopic signal contaminated by the pres-ence of a companion that would appear blended with the primary star in the spectrograph fiber? All background objects detectedin the imaging survey are far enough from their stars and do notenter the fiber ( / (cid:13) ∼ (cid:48)(cid:48) ) used in the velocimetric survey. Thus,they do not pollute the radial velocity signal. On the contrary, thecompanion candidate to HD 43940 (ambiguous case) probablycontribute to the velocimetric signal. Six of the seven comov-ing companion detected, HD 14943B ( ρ = . (cid:48)(cid:48) ρ = . (cid:48)(cid:48) ι Crt B ( ρ = . (cid:48)(cid:48) ρ = . (cid:48)(cid:48) o Gru B( ρ = . (cid:48)(cid:48) o Gru B probably marginallypollutes the velocimetric signal depending on seeing conditions.However, the contribution in terms of flux is generally limitedgiven the flux ratios between the stars and the companions.On the other hand, are the radial velocity variations a resultfrom the gravitational perturbations of the imaged companions?Indeed, all the 6 stars for which we imaged and confirmed a com-panion are reported bearing velocimetric variations in Lagrangeet al. (2009). Radial velocity curves (Lagrange et al. 2009) in-dicate that two of them (HD 41742 and o Gru) belong to mul-tiple systems; this includes a spectroscopic binary surroundedby a wider-separation companion, detected only with adaptiveoptics. The imaged companions are not responsible in thesecases for the radial velocity variations. Meanwhile, the con-firmed companions around the other 4 stars partly or totally con-tribute to the radial velocity variations observed by Lagrange etal. (2009) as well as in new unpublished data (Lagrange et al.in preparation). In the ambiguous cases of HD 43940 and η Hor,the companion candidates could contribute to the observed radialvelocity variations, if they are bound objects.In any case, this shows that the radial velocity curves of atleast 6 out of 35 stars followed spectroscopically (17%) may bea ff ected by stellar multiplicity.
6. Conclusions
We have presented a deep imaging near-infrared survey of 38austral early-type main sequence stars, searching for low-masscompanions at intermediate distances. The overall detection per-formances obtained for this survey, mainly performed in the Kband at 2.2 µ m, allowed us to probe the stellar environmentsat separations > . (cid:48)(cid:48) < .
05. Farther from thestars ( > (cid:48)(cid:48) ), we achieved limiting contrasts between 10 − and10 − using coronagraphy. Although known binary stars were ex-cluded from the survey, we detected 41 companion candidatesaround 23 stars. Using multi-epoch observations, we were ableto determine that 8 candidates are actually comoving compan-ions around 7 stars of the sample, while the other candidates aremainly background contaminants. The comoving companionsare low-mass stars with masses ranging from 0.13 to 0.8 M (cid:12) .A comoving binary system consisting of 2 M / L dwarfs imagedaround HD 49095, with an almost fully resolved orbit, will beof particular interest for the calibration of low-mass star evo-lution models. The comoving companions are detected at pro-jected separations ranging from 11.3 to 159 au .At least 16% of the sample stars, with spectral types rang-ing from F7 to B9, therefore host a low-mass star companion.This figure is likely a lower limit to the real multiplicity frac-tion of early-type stars within this separation range. In addition,all the detected and confirmed companions have projected sepa-rations below 200 au : this range matches the physical projectedseparations of giant planets recently imaged in discs surrounding In the case of HD 41742, there is a 4th companion lying out of thefield of view of this survey.
A-type stars Fomalhaut and HR 8799 (Kalas et al. 2008; Maroiset al. 2008). It is certainly tantalizing to interpret the emergingpopulation of such massive planets as the low-mass tail of the bi-nary distribution of early-type stars. This would imply that suchplanets are formed by gravitational instability rather than coreaccretion like closer-in planets. However, further evidence fromlarger, less-biased surveys, are still required to assess this ques-tion in a statistical and quantitative fashion.Most stars of our the sample are also included in spectro-scopic surveys (Lagrange et al. 2009) seeking for radial velocityvariations. Crossing the results of both surveys show that at least17% of the radial velocity curves may be a ff ected by companionsdetected with adaptive optics.We finally emphasize that tight stellar systems with a brightearly-type primary component, such as the ones presented inthis study, will be excellent calibrators for next-generationplanet imagers such as VLT / SPHERE or interferometers likeVLT / Gravity, designed for high-precision astrometry withinsmall field-of-views.
Acknowledgements.
We are grateful to the referee J. Carson, for havingpromptly provided a useful report on this work. These results have extensivelymade use of the SIMBAD and VizieR databases, operated at CDS, Strasbourg,France. This publication makes use of data products from the Two Micron AllSky Survey, which is a joint project of the University of Massachusetts andthe Infrared Processing and Analysis Center / California Institute of Technology,funded by the National Aeronautics and Space Administration and the NationalScience Foundation. This research has made use of the Washington Double StarCatalogue maintained at the U.S. Naval Observatory. AML, DE, and GC ac-knowledge financial support from the French National Research Agency (ANRcontract NT05-44463). DE also acknowledges support from the Centre Nationald’ ´Etudes Spatiales (CNES).
References
Bara ff e, I., Chabrier, G., Allard, F., Hauschildt, P. H. 1998, A&A, 337, 403Bevington, P. R., & Robinson, K. D. 2003, Data Reduction and Error Analysisfor the Physical Sciences (Boston: McGraw-Hill), 3rd ed.Boccaletti, A., Chauvin, G., Baudoz, P., Beuzit, J.-L. 2007, A&A, 482, 939Bonfils, X., Mayor, M., Delfosse, X., et al. 2007, A&A, 474, 293Boss, A. P. 1997, Science, 276, 1836Boss, A. P. 2006, ApJ, 643, 501Bouchy, F., H´ebrard, G., Udry, S., et al. 2009, A&A, 505, 853Chauvin, G., Lagrange, A.-M., Dumas, C., et al. 2004, A&A, 425, L29Chauvin, G., Lagrange, A.-M., Dumas, C., et al. 2005, A&A, 438, L25Chauvin, G., Lagrange, A.-M., Lacombe, F., et al. 2005, A&A, 430, 1027Chauvin, G., Lagrange, A.-M., Zuckerman, B., et al. 2005, A&A, 438, L29Chauvin, G., Lagrange, A.-M., Udry, S., et al. 2006, A&A, 456, 1165Chauvin, G., Lagrange, A.-M., Bonavita, M., et al. 2010, A&A, 509, A52Cox, A. N. 2000, Allen’s Astrophysical Quantities (New York: AIP Press;Springer), 4th ed.Crida, A., Masset, F., Morbidelli, A. 2009, ApJ, 705, L148Devillard, N. 1997, The Messenger, 87, 19Desidera, S. & Barbieri, M. 2007, A&A, 462, 345Desort, M., Lagrange, A.-M., Galland, F., et al. 2008, A&A, 491, 883Dodson-Robinson, S. E., Veras, D., Ford, E. B., Beichman, C. A. 2009, ApJ,707, 79Doyon, R., Nadeau, D., Vall´ee, P., et al. 1998, SPIE, 3354, 760Ducati, J. R., Bevilacqua, C. M., Rembold, S. B., Daiana, R. 2001, ApJ, 558, 309Duchˆene, G., McCabe, C., Pinte, C., et al. 2010, ApJ, 709, L114Eggen, O. J. 1991, AJ, 102, 1826Eggen, O. J. 1991, AJ, 102, 2028Eggen, O. J. 1998, AJ, 116, 284Eggenberger, A., Udry, S., Chauvin, G., et al. 2007, A&A, 474, 273Eggenberger, A., Udry, S., Mayor, M. 2004, A&A, 417, 353Frankowski, A., Jancart, S., Jorissen, A. 2007, A&A, 464, 377Gray, R. O., Corbally, C. J., Garrison, R. F. et al. 2006, AJ, 132, 161Grosbøl, P. J. 1978, A&AS, 32, 409Goldin, A. & Makarov, V. V. 2007, ApJS, 173, 137Hauck, B. 1986, A&A, 155, 371Hinkley, S., Oppenheimer, B. R., Brenner, D., et al. 2010, ApJ, 712, 421Hubrig, S., Le Mignant, D., North, P., Krautter, J. 2001, A&A, 372, 152Holmberg, J., Nordstr¨om, B., Andersen, J. 2009, A&A, 501, 941 hrenreich et al.: Deep imaging of close companions to A–F stars 13 Huang, W. & Gies, D. R. 2008, ApJ, 683, 1045Ida, S. & Lin, D. N. C. 2004, ApJ, 616, 567Ida, S. & Lin, D. N. C. 2005, ApJ, 626, 1045Kalas, P., Graham, J. R., Chiang, E., et al. 2008, Science, 322, 1345Kennedy, G. M. & Kenyon, S. J. 2007, Ap&SS, 311, 9King, J. R., Villarreal, A. R., Soderblom, D. R., Gulliver, A. F. & Adelman, S. J.2003, AJ, 125, 1980Kratter, K. M., Murray-Clay, R. A., Youdin, A. N. 2010, ApJ, 710, 1375Lagrange, A.-M., Bonnefoy, M., Chauvin, G., et al. 2010, Science, accepted(arXiv:1006.3314)Lagrange, A.-M., Desort, M., Galland, F., Udry, S., Mayor, M. 2009, A&A, 495,335Lagrange, A.-M., Gratadour, D., Chauvin, G. et al. 2009, A&A, 493, L21Lagrange, A.-M., Kasper, M. Boccaletti, A. et al. 2009, A&A, 506, 927Laughlin, G., Bodenheimer, P., Adams, F. C. 2004, ApJ, 612, L73Lenzen, R., Hartung, M., Brandner, W., et al. 2003, SPIE, 4841, 944Lodato, G., Delgado-Donate, E., Clarke, C. J. 2005, MNRAS, 364, L91Lovis, C. & Mayor, M. 2007, A&A, 472, 657Lowrance, P. J., McCarthy, C, Becklin, E. E., et al. 1999, ApJ, 512, L69Lowrance, P. J., Schneider, G., Kirkpatrick, J. D., et al. 2000, ApJ, 541, 390Luhman, K. L. & Jayawardhana, R. 2002, ApJ, 566, 1132Mannings, V. & Barlow, M. J. 1998, ApJ, 497, 330Marois, C., Macintosh, B., Barman, T., et al. 2008, Science, 322, 1348McCaughrean, M. J. & Stau ff er, J. R. 1994, AJ, 108, 1382Morales, F. Y., Werner, M. W., Bryden, G., et al. 2009, ApJ, 699, 1067Mordasini, C., Alibert, Y., Benz, W., Naef, D. 2009, A&A, 501, 1161Mugrauer, M., Seifahrt, A., Neuh¨auser, R. 2007, MNRAS, 378, 1328Neuh¨auser, R., G¨unther, E. W., Alves, J., et al. 2003, AN, 324, 535Makarov, V. V. 2007, ApJS, 169, 105Makarov, V. V. & Kaplan, G. H. 2005, AJ, 129, 2420Mayer, L., Wadsley, J., Quinn, T., Stadel, J. 2005, MNRAS, 363, 641Mayor, M., Bonfils, X., Forveille, T., et al. 2009, A&A, 507, 487Mayor, M., Udry, S., Lovis, C., et al. 2009, A&A, 493, 639Moshir, M. 1989, Explanatory supplement to the IRAS
Faint Source Survey,Infrared Processing and Analysis Center, Pasadena, CA, USANakajima, T., Oppenheimer, B. R., Kulkarni, S. R., et al. 1995, Nature, 378, 463Patience, J., White, R. J., Ghez, A. M., et al. 2002, ApJ, 581, 654Pollack, J. B., Hubickyj, O., Bodenheimer, P., et al. 1996, Icarus, 124, 62Pourbaix, D., Tokovinin, A. A., Batten, A. H., et al. 2004, A&A, 424, 727Plavchan, P., Werner, M. W., Chen, C. H., et al. 2009, A&A, 698, 1068Rafikov, R. R. 2005, ApJ, 621, L69Rieke, G. H., Su, K. Y. L., Stansberry, J. A., et al. 2005, ApJ, 620, 1010Rigaut, F., Salmon, D., Arsenault, R., et al. 1998, PASP, 110, 152Rivera, E. J. & Lissauer, J. J. 2000, ApJ, 530, 454Rousset, G., Lacombe, F., Puget, P., et al. 2003, SPIE, 4839, 140Santos, N. C., Israelian, G., Mayor, M. 2001, A&A, 373, 1019Santos, N. C., Israelian, G., Mayor, M. 2004, A&A, 415, 1153Skrutskie, M. F., Cutri, R. M., Stiening, R. et al. 2006, AJ, 131, 1163Su, K. Y. L., Rieke, G. H., Stansberry, J. A. et al. 2006, ApJ, 653, 675Suchkov, A. A., Makarov, V. V., Voges, W. 2003, ApJ, 595, 1206Sumi, T., Bennett, D. P., Bond, I. A. et al. 2010, ApJ, 710, 1641Torres, G. 1986, A&AS, 64, 105Tokovinin, A. A. 1997, A&AS, 124, 75V´eran, J.-P. 1997, PhD Thesis, ´Ecole Nationale Sup´erieure desT´el´ecommunications, FranceZucker, S., & Mazeh, T. 2002, ApJ, 568, L113Zuckerman, B., Song, I., Webb, R. A. 2001, ApJ, 559, 388
Fig. 4.
Examples of the status determination of companion candidates around HD 14943 (left) and η Tuc (HD 224392; right), from astrometricmeasurements at two di ff erent epochs. The positions on the sky are reported in the upper panel: the candidates were first imaged on 2005-11-07 (filled black circles), then on 2008-08-20 (filled red and blue circles; there are two measurements in the case of HD 14943). The positionspredicted for the candidates on 2008-08-20 if they were background objects from the field are indicated by the empty red and blue circles, andtheir trajectories between the two epochs figured by the black curves, which take into account the stellar proper motions and the parallacticmotions. The middle and lower panels show the evolutions of the projected separations and the position angles, respectively, with time. One-sigmaerror bars are represented. A zero or near-zero evolution of the separation and position angle with time is indicative of a physical association. Thecandidate at the left is bound to its star with indicative probabilities P bkg =
0% and P cmv = .
7% (following the calculations of Sect. 4.2), whilethe candidate at the right is a background contaminant, with P bkg = .
0% and P cmv = Fig. 5.
Example of 6- σ detection limits reached for a K = . × × fields around the star, respectively.The hatched area on the coronagraphic image represents the occulting mask with a diameter of 0 . (cid:48)(cid:48)
7. Classic 1-dimensional 6- σ detection limits(right panel) are extracted from the direct image detection limit map (plain line) and the coronagraphy detection limit map (dotted line) as ex-plained in Sect. 4.3. Mass limits (red lines) are obtained by interpolating Bara ff e et al.’s (1998) evolutionary model for a given stellar age (here,700 Myr). Fig. 6.
Same as Fig. 5 for a K = . ∼ . Fig. 9.
Stars with companion candidates confirmed through multi-epoch observations. North is up and east is to the left. Counts on the detector aredisplayed with a logarithmic scale. h r e n r e i c h e t a l . : D ee p i m a g i ngo f c l o s ec o m p a n i on s t o A – F s t a r s α δ type V J H K ref. code ref. code ref.(mas yr − ) (pc) (mag) (mag) (mag) (mag) (Myr)1 HD 2834 140.74 19.47 A0V 52.7 4.8 4.7 4.8 4.7 600–800 13 V:SB 22 HD 3003 β Tuc 86.15 -49.85 A0V 46.5 5.1 5.1 5.2 5.0 < ∼
40 18 K:B 4 C 23 HD 4293 -89.06 -96.11 F0V 66.6 5.9 5.4 5.3 5.2 10–100 19 C 24 HD 10939 126.45 59.04 A1V 57.0 5.0 5.0 5.0 5.0 320 205 HD 14943 22.09 66.01 A5V 61.3 5.9 5.6 5.5 5.4 850 1 N:B V 26 HD 16555 η Hor 111.29 1.05 A6V 44.5 5.3 4.6 4.6 4.5 50–350 21 V 37 HD 16754 s Eri 92.11 -15.07 A1V 39.8 4.8 ∼
350 22 K:B 48 HD 19545 65.72 -20.57 A5V 57.6 6.2 5.9 5.9 5.8 ∼
27 23 V:SB 29 HD 20888 57.15 13.07 A3V 58.0 6.0 5.8 5.8 5.7 10–40 510 HD 21882 -59.48 -1.45 A5V 62.3 5.8 5.4 5.2 5.2 C 211 HD 29875 α Cae -141.18 -74.95 F2V 20.1 4.5 3.9 3.7 3.7 1 000 6 K:B 4 V 212 HD 29992 46.9 193.14 F3V 27.7 5.1 4.7 4.3 4.1 1 600 6 V 213 HD 31746 100.99 68.25 F5V 30.9 6.1 5.3 5.1 5.0 1 400 6 V 214 HD 32743 η Pic -43.96 27.18 F5V 26.2 5.4 4.8 4.5 4.4 1 400 6 K:B 4 V 215 HD 41742 -79.98 254.28 F4V 27.0 5.9 5.1 4.9 4.7 3 700 6 K:T 7 V:SB 216 HD 43940 -5.51 -22.63 A2V 62.1 5.9 5.6 5.5 5.4 N:B C 217 HD 49095 -204.91 -304.12 F6.5V 24.3 5.9 5.0 4.8 4.7 2 000 6 K:B,N:T 8 V 218 HD 50445 -40.47 -59.14 A3V 55.1 5.9 5.6 5.5 5.5 N:M C 219 HD 68456 -155.04 -297.68 F6V 21.4 4.8 3.9 3.6 3.7 2 500 6 K:B 12 V:SB 220 HD 75171 -63.56 104.15 A4V 60.6 6.0 5.6 5.5 5.5 800 13 V 221 HD 91889 268.49 -672.31 F7V 24.6 5.7 4.9 4.5 4.3 4 100 6 V 222 HD 101198 ι Crt 99.62 126.17 F6.5V 27.0 5.5 4.8 4.3 4.1 4 800 6 K:B 4 V 223 HD 112934 -118.89 -59.5 A9V 54.9 6.6 6.0 5.9 5.8 2 000 6 V:SB 224 HD 116568 66 Vir 157.73 -36.24 F3.5V 30.1 5.8 5.3 4.9 4.7 2 000 6 V:SB 225 HD 153363 26 Oph 49.14 -54.89 F3V 33.3 5.7 5.0 4.8 4.7 1 700 6 K:B 8,14 V 226 HD 158094 δ Ara -53.65 -99.37 B8V 57.4 3.6 3.7 3.7 3.7 125 15 K:B 4 V:B 227 HD 177756 λ Aql -19.68 -90.37 B9V 38.4 3.4 3.5 3.5 3.6 90 24 V:B 228 HD 186543 ν Tel 92.4 -137.4 A7III-IV 52.1 5.3 5.0 5.0 4.9 250 25 V 229 HD 197692 ψ Cap -51.38 -156.66 F5V 14.7 4.2 3.4 3.1 3.1 1 400 6 V 230 HD 200761 θ Cap 79.64 -61.64 A1V 48.5 4.1 4.4 4.3 4.1 V:B 231 HD 209819 ι Aqr 40.45 -57.16 B8V 52.9 4.3 4.4 4.6 4.4 30–60 26 V:B 232 HD 213398 β PsA 59.64 -18.7 A0V 45.5 4.3 4.3 4.3 4.3 240 24 K:B 4 C 233 HD 216385 σ Peg A 521.86 44.1 F7IV 26.8 5.2 4.2 3.9 3.9 2 700 6 K:B 16 V 934 HD 216627 δ Aqr -44.08 -24.81 A3V 48.9 3.3 3.3 3.2 3.2 300 13,17 K:B 12 V:SB 235 HD 219482 176 -26.12 F6V 20.6 5.7 5.1 4.6 4.4 3 700 6 V 236 HD 220729 o Gru 33.16 130.07 F4V 32.0 5.5 4.9 4.7 4.5 1 100 6 N:B V:SB 237 HD 223011 103.11 -36.01 A7III-IV 65.3 6.3 5.9 5.9 5.8 V 238 HD 224392 η Tuc 78.86 -61.1 A1V 48.7 5.0 4.9 4.9 4.8 10–40 18 C 2 E h r e n r e i c h e t a l . : D ee p i m a g i ngo f c l o s ec o m p a n i on s t o A – F s t a r s α δ type V J H K ref. code ref. code ref.(mas yr − ) (pc) (mag) (mag) (mag) (mag) (Myr) Table 1.
Star sample for the southern survey. Infrared J , H , and K magnitudes are extracted from 2MASS (Skrutskie et al. 2006). The ‘RV’ columnindicates the radial velocimetry status (C: constant, V: variable, with possible source of variations being SB: spectral binary, A: magnetic activity,Pu: pulsations, Pl: planets, D: drift) as determined by Lagrange et al. 2009 with HARPS or as relying on SOPHIE data (Desort et al., personalcommunication). The column ‘Multiplicity’ indicates apparent (or comoving) binary (B), ternary (T), or more-component (M) systems, whetherit is new (N) or known before this survey (K). References are given below the table.References: (1) Su et al. 2006; (2) HARPS survey, Lagrange et al. 2009; (3) HARPS survey, unpublished data; (4) CCDM II, Vizier Online DataCatalogue I / / B , Pourbaix et al. 2004; (11) Desort et al. 2008;(12) Goldin & Makarov 2007; (13) Eggen 1998; (14) proper motion binary, Frankowski et al. 2007; (15) Rieke et al. 2005; (16) The WashingtonVisual Double Star Catalogue, Mason et al. 2001; (17) King et al. 2003; (18) member of the Tucana association, Zuckermann, Song & Webb 2001;(19) HD 4293 is classified as a F0V but has colours corresponding to a F0III giant (Hauck 1986) with typical age between 10 and 100 Myr (Eggen1991a); (20) Morales et al. 2009; (21) Plavchan et al. 2009; (22) Gray et al. 2006; (23) Makarov 2007 suggested that HD 19545 and BO Microscopii(‘Speedy Mic’) were formerly part of the same system within the Tucana-Horologium association; (24) Grosbøl 1978; (25) member of the IC 2391supercluster (Eggen 1991b); (26) a field B star, Huang & Gies 2008hrenreich et al.: Deep imaging of close companions to A–F stars 19 / Instrument Camera Filter Mode CC / NaCo S27 K s D + C 02 HD 3003 II 2005-11-07 VLT / NaCo S27 K s D + C 03 HD 4293 II 2005-11-08 VLT / NaCo S27 K s D + C 04 HD 10939 II 2005-11-06 VLT / NaCo S27 K s D + C 05 HD 14943 II 2005-11-07 VLT / NaCo S27 K s D + C 1V 2008-08-20 VLT / NaCo S13,S27 K s D 16 HD 16555 II 2005-11-06 VLT / NaCo S27 J,K s D + C 1V 2008-08-20 VLT / NaCo S27 K s D + C 17 HD 16754 II 2005-11-07 VLT / NaCo S27 K s D + C 08 HD 19545** V 2008-08-20 VLT / NaCo S27 K s D 09 HD 20888* II 2005-11-07 VLT / NaCo S27 K s D + C 110 HD 21882 II 2005-11-06 VLT / NaCo S13,S27 J,K s D + C 1V 2008-08-21 VLT / NaCo S27 K s D + C 111 HD 29875* II 2005-11-08 VLT / NaCo S27 K s D + C 212 HD 29992 V 2008-08-20 VLT / NaCo S27 K s D + C 013 HD 31746 II 2005-11-07 VLT / NaCo S27 K s D + C 014 HD 32743* II 2005-11-07 VLT / NaCo S27 J,K s D + C 415 HD 41742 II 2005-11-06 VLT / NaCo S27 K s D + C 3IV 2007-11-17 CFHT / PUEO KIR Br γ D + S 3V 2008-08-21 VLT / NaCo S27 K s D + C 3VI 2009-04-25 VLT / NaCo S27 K s D 116 HD 43940 II 2005-11-08 VLT / NaCo S27 K s D + C 1V 2008-08-20 VLT / NaCo S13,S27 H,K s D 1?VI 2009-04-26 VLT / NaCo S13 J D 017 HD 49095 II 2005-11-07 VLT / NaCo S27 K s D + C 3III 2007-01-27 CFHT / PUEO KIR K’,Br γ D + S 2IV 2007-11-16 CFHT / PUEO KIR Fe ii D 2VI 2009-04-26 VLT / NaCo S13 H D 218 HD 50445 II 2005-11-06 VLT / NaCo S27 K s D + C 5VI 2009-04-26 VLT / NaCo S27 K s C*** 519 HD 68456 II 2005-11-07 VLT / NaCo S27 K s D + C 1VI 2009-04-26 VLT / NaCo S27 K s D + C 120 HD 75171* II 2005-11-08 VLT / NaCo S27 K s D + C 221 HD 91889 I 2005-01-27 CFHT / PUEO KIR Br γ D + S 1VI 2009-04-26 VLT / NaCo S27 K s D + C 122 HD 101198 IV 2007-11-16 CFHT / PUEO KIR Fe ii D 1VI 2009-04-26 VLT / NaCo S13,S27 H,K s D 123 HD 112934** III 2007-01-28 CFHT / PUEO KIR K D + S 0?V 2008-08-20 VLT / NaCo S27 K s D 024 HD 116568** V 2008-08-20 VLT / NaCo S27 K s D 025 HD 153363 V 2008-08-20 VLT / NaCo S27 K s D + C MVI 2009-04-26 VLT / NaCo S13 J,H D 226 HD 158094 V 2008-08-21 VLT / NaCo S27 K s D 0VII 2009-08-27 VLT / NaCo S27 K s D + C 327 HD 177756** V 2008-08-21 VLT / NaCo S27 K s D 1VI 2009-04-26 VLT / NaCo S13,S27 H,NB . D 128 HD 186543 V 2008-08-20 VLT / NaCo S27 K s D + C 029 HD 197692 II 2005-11-07 VLT / NaCo S27 K s D + C 1V 2008-08-21 VLT / NaCo S27 K s D + C 130 HD 200761 V 2008-08-21 VLT / NaCo S27 K s D 0VII 2009-08-27 VLT / NaCo S27 K s D + C 031 HD 209819 V 2008-08-21 VLT / NaCo S27 K s D 0VII 2009-08-27 VLT / NaCo S27 K s D + C 032 HD 213398 II 2005-11-06 VLT / NaCo S27 J,K s D + C 1V 2008-08-20 VLT / NaCo S27 K s D + C 133 HD 216385 IV 2007-11-16 CFHT / PUEO KIR Br γ ,Fe ii D + S 1V 2008-08-21 VLT / NaCo S27 K s D + C 134 HD 216627 V 2008-08-21 VLT / NaCo S27 K s D 0VII 2009-08-27 VLT / NaCo S27 NB . D + C 035 HD 219482 V 2008-08-21 VLT / NaCo S27 K s D + C 1VI 2009-04-26 VLT / NaCo S27 K s D + C 136 HD 220729 II 2005-11-07 VLT / NaCo S27 K s D + C 1V 2008-08-20 VLT / NaCo S13,S27 H,K s D 137 HD 223011** V 2008-08-20 VLT / NaCo S27 K s D 0?38 HD 224392 II 2005-11-07 VLT / NaCo S27 K s D + C 2V 2008-08-21 VLT / NaCo S27 K s D + C 2
Table 3.
Observing setups, dates, and number of companion candidates observed (‘M’ stands for many). For each observing date, we precise theobserving mode (D: direct imaging, C: coronagraphic imaging, S: saturated direct imaging), the filter(s) (NaCo filters: J, H, K s , NB . ; PUEOfilters: K’, Fe ii , Br γ ), and camera(s) used (NaCo cameras: S27, S13; PUEO camera: KIR). The last column ‘Companionship’ indicates thepresence of companion candidate(s) (CC) or bound companion(s) (BC) when this has been established at several epochs. The epochs II, V, VI,and VII correspond to ESO observing programmes 076.C-0270(A), 081.C-0653(A), 083.C-0151(B), and 083.C-0151(A), respectively. Epochs I,III, and IV correspond to CFHT programmes . . .*No second-epoch observation was recorded.**No coronagraphic image was recorded. E h r e n r e i c h e t a l . : D ee p i m a g i ngo f c l o s ec o m p a n i on s t o A – F s t a r s × NDIT × NEXP Projected separation Position angle ∆ m ± . ρ ( (cid:48)(cid:48) ) θ ( ◦ ) (mag)5 HD014943 1 2005-11-07 Ks S27 D 1 . × ×
10 2 . ± . . ± .
54 5 .
10 Comoving5 HD014943 1 2008-08-20 Ks S27 D 2 . × ×
10 2 . ± . . ± .
53 5 .
10 Comoving5 HD014943 1 2008-08-20 Ks S13 D 5 . × ×
10 2 . ± . . ± .
41 5 .
20 Comoving6 HD016555 1 2005-11-06 Ks S27 C 2 . × ×
20 4 . ± . − . ± .
77 9 .
00 Ambiguous6 HD016555 1 2008-08-20 Ks S27 C 2 . × ×
10 4 . ± . − . ± .
44 10 . . × ×
10 4 . ± . − . ± .
44 10 . . × ×
10 4 . ± . . ± .
19 3 .
80 Undefined10 HD021882 1 2005-11-06 Ks S27 C 6 . × ×
18 5 . ± . − . ± .
18 11 . . × ×
10 5 . ± . − . ± .
17 10 . . × ×
10 7 . ± . . ± .
61 3 .
90 Undefined11 HD029875 2 2005-11-08 Ks S27 D 0 . × ×
10 7 . ± . . ± .
43 4 .
20 Undefined14 HD032743 1 2005-11-07 Ks S27 D 0 . × ×
10 12 . ± . − . ± .
12 7 .
40 Undefined14 HD032743 2 2005-11-07 Ks S27 C 6 . × ×
18 12 . ± . − . ± .
12 9 .
10 Undefined14 HD032743 3 2005-11-07 Ks S27 C 6 . × ×
18 15 . ± . − . ± .
12 11 . . × ×
18 15 . ± . − . ± .
12 12 . . × ×
10 5 . ± . − . ± .
45 1 .
80 Comoving15 HD041742 1 2007-11-17 BrG KIR D 1 . × ×
25 5 . ± . − . ± .
37 1 .
70 Comoving15 HD041742 1 2008-08-21 Ks S27 D 0 . × ×
10 5 . ± . − . ± .
40 1 .
70 Comoving15 HD041742 1 2009-04-25 Ks S27 D 0 . × ×
14 5 . ± . − . ± .
35 1 .
50 Comoving15 HD041742 2 2005-11-06 Ks S27 C 4 . × × . ± . . ± .
14 8 .
10 Background15 HD041742 2 2007-11-17 BrG KIR S 40 . × ×
25 7 . ± . . ± .
13 10 . . × ×
10 7 . ± . . ± .
13 7 .
50 Background15 HD041742 3 2005-11-06 Ks S27 C 4 . × × . ± . − . ± .
14 8 .
80 Background15 HD041742 3 2007-11-17 BrG KIR S 40 . × ×
25 7 . ± . − . ± .
13 10 . . × ×
10 7 . ± . − . ± .
13 7 .
90 Background16 HD043940 1 2005-11-08 Ks S27 D 1 . × ×
10 0 . ± . − . ± .
86 2 .
10 Ambiguous16 HD043940 1 2008-08-20 Ks S27 D 2 . × ×
10 0 . ± . − . ± . .
10 Ambiguous16 HD043940 1 2008-08-20 H S13 D 5 . × ×
10 0 . ± . − . ± . .
40 Ambiguous17 HD049095 1 2005-11-07 Ks S27 D 0 . × ×
10 1 . ± . − . ± .
42 5 .
40 Comoving17 HD049095 1 2007-01-27 BrG KIR D 3 . × ×
35 1 . ± . − . ± .
55 4 .
80 Comoving17 HD049095 1 2007-11-16 FeII KIR D 1 . × ×
25 1 . ± . − . ± .
47 4 .
70 Comoving17 HD049095 1 2009-04-26 H S13 D 0 . × ×
21 1 . ± . − . ± .
00 5 .
60 Comoving17 HD049095 2 2005-11-07 Ks S27 D 0 . × ×
10 1 . ± . − . ± .
34 5 .
60 Comoving17 HD049095 2 2007-01-27 BrG KIR D 3 . × ×
35 1 . ± . − . ± .
53 4 .
90 Comoving17 HD049095 2 2007-11-16 FeII KIR D 1 . × ×
25 1 . ± . − . ± .
41 4 .
90 Comoving17 HD049095 2 2009-04-26 H S13 D 0 . × ×
21 1 . ± . − . ± .
82 6 .
10 Comoving18 HD050445 1 2005-11-06 Ks S27 C 5 . × ×
20 12 . ± . . ± .
12 8 .
30 Background18 HD050445 1 2009-04-26 Ks S27 C 5 . × ×
10 12 . ± . . ± .
08 8 .
50 Background18 HD050445 2 2005-11-06 Ks S27 C 5 . × ×
20 15 . ± . . ± .
18 8 .
00 Background18 HD050445 2 2009-04-26 Ks S27 C 5 . × ×
10 14 . ± . . ± .
12 7 .
90 Background18 HD050445 3 2005-11-06 Ks S27 C 5 . × ×
20 14 . ± . . ± .
16 10 . . × ×
10 14 . ± . . ± .
11 10 . . × ×
20 7 . ± . . ± .
14 11 . . × ×
10 8 . ± . . ± .
11 11 . . × ×
20 10 . ± . − . ± .
13 12 . . × ×
10 10 . ± . − . ± .
10 12 . . × ×
20 5 . ± . . ± .
18 10 . . × ×
20 6 . ± . . ± .
13 10 . h r e n r e i c h e t a l . : D ee p i m a g i ngo f c l o s ec o m p a n i on s t o A – F s t a r s × NDIT × NEXP Projected separation Position angle ∆ m ± . ρ ( (cid:48)(cid:48) ) θ ( ◦ ) (mag)20 HD075171 1 2005-11-08 Ks S27 C 10 . × ×
20 4 . ± . − . ± .
20 10 . . × ×
20 14 . ± . − . ± .
13 11 . . × ×
15 9 . ± . − . ± .
16 12 . . × ×
10 11 . ± . − . ± .
08 10 . . × ×
25 0 . ± . − . ± .
16 3 .
10 Comoving22 HD101198 1 2009-04-26 H S13 D 0 . × ×
12 0 . ± . − . ± .
55 3 .
10 Comoving22 HD101198 1 2009-04-26 Ks S27 D 0 . × ×
12 0 . ± . − . ± .
86 3 .
00 Comoving25 HD153363 1 2008-08-20 Ks S27 D 2 . × ×
10 0 . ± . − . ± .
32 2 .
00 Comoving25 HD153363 1 2009-04-26 H S13 D 0 . × ×
60 0 . ± . − . ± .
11 2 .
70 Comoving25 HD153363 1 2009-04-26 J S13 D 0 . × ×
20 0 . ± . − . ± .
08 3 .
10 Comoving25 HD153363 2 2008-08-20 Ks S27 D 2 . × ×
10 8 . ± . − . ± .
14 8 .
00 Background25 HD153363 2 2009-04-26 H S13 D 0 . × ×
60 8 . ± . − . ± .
39 7 .
60 Background26 HD158094 1 2009-08-27 Ks S27 C 0 . × ×
10 3 . ± . . ± .
25 10 . . × ×
10 8 . ± . . ± .
10 9 .
30 Undefined26 HD158094 3 2009-08-27 Ks S27 C 0 . × ×
10 12 . ± . . ± .
08 8 .
20 Undefined27 HD177756 1 2008-08-21 Ks S27 D 1 . × ×
10 12 . ± . . ± .
10 4 .
00 Background27 HD177756 1 2009-04-26 Ks S27 D 1 . × × . ± . . ± .
08 3 .
80 Background29 HD197692 1 2005-11-07 Ks S27 C 0 . × ×
20 11 . ± . . ± .
13 12 . . × ×
10 11 . ± . . ± .
11 12 . . × ×
20 9 . ± . − . ± .
14 10 . . × ×
20 9 . ± . − . ± .
14 9 .
50 Background32 HD213398 1 2008-08-20 Ks S27 C 2 . × ×
10 9 . ± . − . ± .
11 10 . . × ×
30 15 . ± . . ± .
07 6 .
30 Background33 HD216385 2 2008-08-21 Ks S27 C 1 . × ×
10 14 . ± . . ± .
18 12 . . × ×
10 10 . ± . − . ± .
12 11 . . × ×
10 10 . ± . − . ± .
10 12 . . × ×
10 0 . ± . . ± .
84 4 .
70 Comoving36 HD220729 1 2008-08-20 Ks S27 D 1 . × ×
10 0 . ± . . ± .
17 4 .
60 Comoving36 HD220729 1 2008-08-20 H S13 D 4 . × ×
10 0 . ± . . ± .
24 4 .
40 Comoving38 HD224392 1 2005-11-07 Ks S27 C 0 . × ×
20 13 . ± . . ± .
12 7 .
50 Background38 HD224392 1 2008-08-21 Ks S27 C 0 . × ×
20 13 . ± . . ± .
11 7 .
40 Background38 HD224392 2 2005-11-07 Ks S27 C 4 . × ×
10 12 . ± . . ± .
13 10 . . × ×
10 12 . ± . . ± .
11 10 . Table 5.
Multi-epoch measurement results for 41 companion candidates detected around 23 stars of the sample. The imaging mode used for eachdetection is indicated: (D) direct imaging, (S) saturated imaging (only on PUEO), (C) coronagraphic imaging (only on NaCO). The companioncandidate status are defined in Sect. 4.2.Notes for some targets:HD 32743 — ρ , θ , and ∆ m of CC ∼
56) are seen using the coronagraph;however, they are all likely background contaminants as the star is close to the Galactic plane ( b = + . ◦◦