BEAST begins: Sample characteristics and survey performance of the B-star Exoplanet Abundance Study
Markus Janson, Vito Squicciarini, Philippe Delorme, Raffaele Gratton, Mickael Bonnefoy, Sabine Reffert, Eric E. Mamajek, Simon C. Eriksson, Arthur Vigan, Maud Langlois, Natalia Engler, Gael Chauvin, Silvano Desidera, Lucio Mayer, Gabriel-Dominique Marleau, Alexander J. Bohn, Matthias Samland, Michael Meyer, Valentina d'Orazi, Thomas Henning, Sascha Quanz, Matthew Kenworthy, Joseph C. Carson
aa r X i v : . [ a s t r o - ph . E P ] J a n Astronomy & Astrophysicsmanuscript no. main © ESO 2021January 7, 2021
BEAST begins: Sample characteristics and survey performance ofthe B-star Exoplanet Abundance Study ⋆ Markus Janson , Vito Squicciarini , , Philippe Delorme , Ra ff aele Gratton , Mickaël Bonnefoy , Sabine Re ff ert , EricE. Mamajek , , Simon C. Eriksson , Arthur Vigan , Maud Langlois , Natalia Engler Gaël Chauvin , , SilvanoDesidera , Lucio Mayer , Gabriel-Dominique Marleau , , , Alexander J. Bohn , Matthias Samland , MichaelMeyer Valentina d’Orazi , Thomas Henning , Sascha Quanz , Matthew Kenworthy , and Joseph C. Carson Institutionen för astronomi, Stockholms Universitet, Stockholm, Swedene-mail: [email protected] Dipartimento di Fisica e Astronomia “Galileo Galilei” Università di Padova, Padova, Italy INAF - Osservatorio Astronomico di Padova, Padova, Italy Univ. Grenoble Alpes, IPAG, Grenoble, France Landessternwarte, Zentrum für Astronomie der Universität Heidelberg, Heidelberg, Germany Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA Department of Physics and Astronomy, University of Rochester, Rochester, NY, USA Aix Marseille Université, CNRS, LAM, Marseille, France CRAL, CNRS, Université Lyon, Saint Genis Laval, France ETH Zurich, Zurich, Switzerland Departemento de Astronomía, Universidad de Chile, Santiago, Chile Center for Theoretical Astrophysics and Cosmology, Institute for Computational Science, University of Zurich, Zurich, Switzer-land Institut für Astronomie und Astrophysik, Eberhard Karls Universität Tübingen, Tübingen, Germany Physikalisches Institut, Universität Bern, Bern, Switzerland Max Planck Institut für Astronomie, Heidelberg, Germany Leiden Observatory, Leiden University, Leiden, The Netherlands Department of Astronomy, University of Michigan, Ann Arbor, MI, USA College of Charleston, Charleston, SC, USAReceived —; accepted —
ABSTRACT
While the occurrence rate of wide giant planets appears to increase with stellar mass at least up through the A-type regime, B-typestars have not been systematically studied in large-scale surveys so far. It therefore remains unclear up to what stellar mass thisoccurrence trend continues. The B-star Exoplanet Abundance Study (BEAST) is a direct imaging survey with the extreme adaptiveoptics instrument SPHERE, targeting 85 B-type stars in the young Scorpius-Centaurus (Sco-Cen) region with the aim to detect giantplanets at wide separations and constrain their occurrence rate and physical properties. The statistical outcome of the survey willhelp determine if and where an upper stellar mass limit for planet formation occurs. In this work, we describe the selection andcharacterization of the BEAST target sample. Particular emphasis is placed on the age of each system, which is a central parameterin interpreting direct imaging observations. We implement a novel scheme for age dating based on kinematic sub-structures withinSco-Cen, which complements and expands upon previous age determinations in the literature. We also present initial results fromthe first epoch observations, including the detections of ten stellar companions, of which six were previously unknown. All planetarycandidates in the survey will need follow up in second epoch observations, which are part of the allocated observational programmeand will be executed in the near future.
Key words.
Planets and satellites: detection – Stars: early-type – Brown dwarfs
1. Introduction
Studying planetary populations in a range of stellar environ-ments is of critical importance for understanding their for-mation and early evolution. High-contrast imaging with adap-tive optics (AO) is a valuable technique for studying the de-mographics of wide giant planets in this context, in virtu-ally any kind of environment. Direct imaging also preferen-tially facilitates the study of young systems, representing an ⋆ Based on observations from the European Southern Observatory,Chile (Programmes 1101.C-0258 and 0103.C-0251). early and potentially pristine stage of the evolution of thesystem. Consequently, a large number of surveys have beenperformed to better understand this population and its dis-tribution in a range environments such as around low-massstars (e.g. Delorme et al. 2012; Bowler et al. 2015; Lannier et al.2016), Sun-like stars (e.g. Lafrenière et al. 2007; Brandt et al.2014; Chauvin et al. 2015), A-type stars (e.g. Vigan et al.2012; Rameau et al. 2013a), binaries (e.g. Bonavita et al. 2016;Asensio-Torres et al. 2018; Hagelberg et al. 2020), disc hosts(e.g. Janson et al. 2013a; Meshkat et al. 2017; Lombart et al.2020), as well as broader surveys covering multiple such
Article number, page 1 of 22 & Aproofs: manuscript no. main demographics (e.g. Nielsen et al. 2019; Desidera et al. 2020).These e ff orts have resulted in the detections of several plan-ets and low-mass substellar companions (e.g. Marois et al.2008; Lagrange et al. 2010; Kuzuhara et al. 2013; Rameau et al.2013b; Macintosh et al. 2015; Chauvin et al. 2017), includingsome that appear to still be undergoing formation (Keppler et al.2018; Ha ff ert et al. 2019; Eriksson et al. 2020). A clear trendthat has emerged from such studies is that giant planets appearto be significantly more common around more massive starsthroughout the M-type to A-type range (e.g. Crepp & Johnson2011; Nielsen et al. 2019; Vigan et al. 2020). The extremes ofthis range are particularly informative in this context, so a crit-ical question is how such relations extend to yet higher masses.Unfortunately, much less is known in the B-type stellar regime.Only one dedicated survey for planets around B-type stars hasbeen performed (Janson et al. 2011), which utilized a previous-generation AO system and featured a modest sample size.In this paper, we outline the B-star Exoplanet AbundanceStudy (BEAST), which is a dedicated survey for planets aroundB-type stars with the SPHERE instrument (Beuzit et al. 2019) atthe Very Large Telescope (VLT). The BEAST study is run as aso-called large programme, spanning several observing periods,and targets 85 stars in the Sco-Cen young stellar association.In parallel, we are also analysing archival data of individual B-stars in Sco-Cen that have been observed in other programmes.This has already led to the detection of the low-mass substellarcircumbinary companion HIP 79098 (AB)b (Janson et al. 2019).In this paper, however, we focus on the large programme thatforms the core of BEAST.While specific B-star surveys have been limited, individuallate B-type objects have been observed in various programmes,which has already resulted in several detections of low-masssubstellar companions (e.g. Lafrenière et al. 2011; Carson et al.2013; Cheetham et al. 2018). This seems to imply that the trendof increasing wide giant planet frequency from M- to A-typestars continues at least up through ∼ B9 types. Meanwhile, planetfrequency as a function of stellar mass has also been examined inthe context of radial velocity (RV) studies. While massive starsare di ffi cult to study with RV on the main sequence (MS) be-cause of their lack of narrow spectral lines, these stars expandand cool in the post-MS evolution resulting in more numerousand narrower lines, thereby making these star much more suit-able for such studies. Radial velocity preferentially covers smallorbital separations, and thus forms an excellent complement todirect imaging for probing planet frequency at a range of separa-tions. Intriguingly, the RV studies show the same trend as directimaging studies with an increasing giant planet frequency as afunction of stellar mass, up to a mass of ∼ M sun (Johnson et al.2010; Re ff ert et al. 2015, Woltho ff et al. in prep). However, inthe mass range 2–3 M sun , the frequency turns over and starts de-creasing with increasing stellar mass (Re ff ert et al. 2015). Inci-dentally, a mass of 2.5 M sun roughly corresponds to a spectraltype of B9 on the MS, and thus marks the approximate transi-tion between the A and B spectral type ranges.The decrease in planet frequency in the 2–3 M sun range inRV studies could be seen either as a formation-related or amigration-related issue. On one hand, discs around massive starsare probably more massive themselves, which probably benefitsplanet formation. Meanwhile, they experience higher doses ofhigh-energy radiation from the central star than planets aroundlower-mass stars, so the disc also potentially dissipates faster. Inrelation to the observed apparent planet abundance turnover at ∼ M sun , in RV studies (Re ff ert et al. 2015) this could be inter- preted as the point at which the negative e ff ects of the dissipationoutweighs the positive e ff ects of an enhanced mass reservoir.On the other hand, giant planets at su ffi ciently close separa-tions to be observable with RV are generally expected to haveformed at larger separations than their current orbit and subse-quently migrated inwards in the disc (e.g. Mordasini et al. 2009).Theoretical work (Kennedy & Kenyon 2008) predicts that thegradual evolution of the protoplanetary disc causes migrationto slow down around more massive stars. Hence, the giantplanet frequency may potentially keep increasing beyond a stel-lar mass of 2 M sun , but simultaneously, the planets may be in-creasingly prohibited from migrating into the inner parts of thesystem, where they could have been detectable with RV. Bothof these possible scenarios make identical predictions for RVsurveys. But they make diametrically opposite predictions fordirect imaging, because the wide giant planet frequency woulddecrease in the 2–3 M sun range if formation is halted, while thefrequency would instead increase if migration is halted. A di-rect imaging survey such as BEAST is required to address thisambiguity.Moreover, specific formation scenarios, such as disc insta-bility, may occur preferentially in the outermost regions of mas-sive discs encircling massive stars (Helled et al. 2014). Accel-erated disc dissipation would not be an issue in this case be-cause disc instability occurs fast, likely in the first 10 yr of thedisc lifetime. Conversely, enhanced stellar irradiation would in-crease the temperature of the inner disc, possibly slowing downor even stopping migration (Rowther & Meru 2020), which canotherwise be very fast in unstable discs (Müller et al. 2018). Asa result, a population of substellar companions formed by discinstability (gas giants as well as brown dwarfs) might be com-pletely absent in RV surveys, while it should be fully accessi-ble to BEAST. The discovery of a coherent excess populationof gas giants and brown dwarfs with wide orbits and a rela-tively top-heavy mass distribution would be a clear indicationin favour of this distribution having been formed via disc in-stability. Early indications of two distinct population of directlyimaged planets / brown dwarfs already exists around lower-massstars (Nielsen et al. 2019; Vigan et al. 2020) and may potentiallybe considerably more clearly distinguished in a survey of mas-sive stars.The BEAST survey is entirely focussed on the Sco-Cen re-gion (de Zeeuw et al. 1999), which is a young ( ∼ ∼ ff ers one of the most favourable trade-o ff s betweenproximity and youth out of all known associations of stars (e.g.Pecaut & Mamajek 2016). Both proximity and youth are impor-tant factors in facilitating detections with direct imaging, butthey are partly mutually exclusive because younger stars arerarer and thus, on average, more distant. For example, there aremore nearby young moving groups than Sco-Cen, but they arenot as young, or not large enough to contain any significant num-ber of B-type stars (e.g. Zuckerman & Song 2004; Gagné et al.2018). Conversely, there are younger associations and clustersthan Sco-Cen, but they are more distant and often exhibit sub-stantial levels of extinction. Sco-Cen is the most nearby coherentregion that o ff ers a large ( > Article number, page 2 of 22. Janson et al.: BEAST survey tical analysis of the full survey. Another crucial benefit to usingthe Sco-Cen region is that it is already heavily targeted for lower-mass (F and A-type) stars in the SHINE survey (Desidera et al.2020), as well as other smaller surveys using SPHERE. Hence,we will be able to statistically compare the BEAST sample witha lower-mass sample in the same region. This comparison e ff ec-tively eliminates the impact of factors such as age and metallic-ity, since they are relatively uniform across the region and thuscleanly isolates stellar mass as the fundamental parameter to testplanet occurrence rate against.
2. BEAST sample
Our initial master list for BEAST target selection consistedof all B-type stars identified as members of Sco-Cen with >
50% probability in a Bayesian kinematic membership studyby Rizzuto et al. (2011). This input list contains 165 potentialtargets. We also added 11 targets that were identified in thede Zeeuw et al. (1999) list of Sco-Cen objects, and for whichthe BANYAN (Gagné et al. 2014, 2018) code provided a > ′′ were removed. The reason for this choice isthat secondaries in this separation range would have a detrimen-tal e ff ect on the contrast performance of SPHERE. Binaries inthe relevant separation range are generally identified from theWashington Double Star (WDS; Mason et al. 2001) survey.(2) Targets that have already been observed by SPHERE withcomparable settings and comparable depth were also removedfrom the sample. This is done to avoid duplications in accor-dance with European Southern Observatories (ESO) policy. Ourlong-term plan is to eventually combine the BEAST survey withall other su ffi ciently deep SPHERE observations of B-type starsin Sco-Cen to yield a full census of the B-star wide giant planetpopulation in this region.(3) Objects with a declination within approximately ± ff erential imaging (ADI) purposes(Marois et al. 2006). Hence, ADI cannot be performed with highe ffi ciency for such targets.Our final BEAST target sample consists of 85 B-type starsthat fulfil all of the constraints for selection. The BEAST surveyis the first to cover a large sample of young B-type stars – acomparison to other direct imaging surveys in terms of spectraltype and age is shown in Fig. 1. In the following, we discuss thedetermination of the detailed characteristics for the targets in thesample. The majority of our sample originates from the Rizzuto et al.(2011) Bayesian analysis of Sco-Cen membership, wherein theobjects were determined as probable members of the association.However, with new astrometry from
Gaia
DR2 (Brown et al.2018) and more sophisticated Bayesian models (Gagné et al.
B0 B5 A0 A5 F0Spectral type10 A ge ( M y r) BEAST SHINE GPIESOther
Fig. 1.
Comparison of BEAST (blue asterisks), SHINE (Desidera et al.2020) (red circles), and GPIES (Nielsen et al. 2019) samples (greencrosses). Also shown as brown triangles are targets from other previ-ous surveys covering early-type stars (Janson et al. 2011; Nielsen et al.2013). To enhance visibility, a Gaussian stochastic term with standarddeviation of 1 Myr was added to all stellar ages in the figure. The mas-sive stellar range covered in BEAST has not been previously systemat-ically probed in direct imaging studies.
Gaia
DR2. A comparison between
Gaia
DR2 and thevan Leeuwen (2007) reduction of
Hipparcos data for the BEASTtargets is shown in Fig. 2.
Gaia has somewhat smaller formal er-ror bars (median of 0.21 mas) than
Hipparcos (median of 0.34mas). However, this does not necessarily translate to a higher ac-curacy, at least not in all cases, because the
Gaia
DR2 reductiononly attempts to fit the parallax and proper motion. Thereforeany photocentre shifts due to binarity within the systems leadto systematic errors in the fitted quantities. The magnitude of thee ff ect depends on the separation and flux ratio of the binary com-ponents. This is particularly relevant for BEAST, whose targetsare early-type and thus have a relatively high multiplicity frac-tion (Duchêne & Kraus 2013). The target stars are also outsideof the brightness range for which the DR2 reduction is optimized( > ff ects. Thescatter around a 1-to-1 relationship in Fig. 2 is 1.28 mas, whichis much larger than the quoted estimated errors of either datasource. We therefore conclude that the astrometric precision formost individual B-type stars in Sco-Cen is probably limited to ∼ Hipparcos or Gaia values are more accurate.A concrete example is the HIP 82514 system, which is aneclipsing binary. As a consequence, the radii and temperaturesof the individual central stellar components can be accuratelydetermined, leading to an unusually reliable photometric dis-tance of ∼
135 pc (Budding et al. 2015). Meanwhile, the distancebased on the
Hipparcos parallax is 154 pc, and the
Gaia
DR2distance is 268 pc. This is a clear indication that the binarity ofHIP 82514 has influenced the astrometric solutions, which may
Article number, page 3 of 22 & Aproofs: manuscript no. main have a particularly strong e ff ect for the Gaia analysis. Very re-cently, an early version of the third
Gaia data release, EDR3,has been made public. We investigated this data release in thecontext of the BEAST sample and found a good general con-sistency on the population level. EDR3 has yet smaller formaluncertainties than DR2, but this data release still only fits forparallax and proper motion, so systematic e ff ects for binaries re-main. Indeed, for HIP 82514, the EDR3 parallax-based distanceis 534 pc, which is much larger than both the DR2 distance andthe photometric distance. Given the large degree of multiplic-ity in our sample, it is not obvious that the increase in precisionwith EDR3 would necessarily reflect an improvement in accu-racy. We thus consistently use DR2 astrometry in this study; thecorresponding distances are shown in Table 5 and in Fig. 3. Later Gaia releases, which will attempt to account for binarity in thesources, should yield significantly better astrometric values interms of both precision and accuracy. G a i a pa r a ll a x ( m a s ) Fig. 2.
Measured parallaxes from
Hipparcos vs.
Gaia for the targets inour sample. The solid diagonal line denotes a 1-to-1 relationship. Thescatter is larger than captured by the estimated individual error bars,implying excess errors from, for example unresolved multiplicity andhigh target brightness.
50 100 150 200 250 300Distance (pc)02468101214 N u m be r o f t a r ge t s Fig. 3.
Histogram of distances for the BEAST sample, from
Gaia whereavailable and from
Hipparcos otherwise.
Multiplicity in the sources can also a ff ect their measuredradial velocities (RVs), causing them to deviate strongly fromthe systemic velocity. In the standard Bayesian models such asBANYAN Σ , the model assumes that the input values correspondto the systemic values. Thus if they are substantially a ff ectedby binary motion, the membership assessment can become sys-tematically erroneous. We attempt to account for such e ff ects bychecking how the BANYAN output is a ff ected if an RV value isincluded in the analysis versus if it is not, and whether the resultschange significantly from the previous analysis based on Hip-parcos astrometry. While most systems show good consistencybetween di ff erent scenarios (as expected), there are some casesin which the outcomes are markedly di ff erent. For example, HIP60855, HIP 65021, and HIP 76126 all had >
50% probabilitiesof membership at the selection phase, but have essentially 0%probabilities in the updated analysis, both when RV data is in-cluded in the analysis and when it is not. Following our discus-sion about systematic uncertainties in the astrometric analysis,membership should not necessarily be categorically excluded insuch cases; however, an individual age estimation then becomesparticularly important.The membership probabilities for each individual target areshown in Table 2. The parallax and proper motion quantities andderived probabilities in the table are generally from
Gaia
DR2astrometry, except for four of the brightest targets which do nothave DR2 astrometry. In these cases,
Hipparcos astrometry fromvan Leeuwen (2007) was used instead.
The age of a target system is a crucial parameter in direct imag-ing studies. In particular, it is a necessary parameter in isochronalanalysis for converting the detected flux from substellar compan-ions into mass estimations, or equivalently, converting upper fluxlimits on non-detection cases into mass detection limits. Massestimations and limits are in turn necessary in statistical studiesfor attempting to interpret the survey results in the context of the-oretical predictions (e.g. Nielsen et al. 2019; Vigan et al. 2020).However, age estimations of individual systems are often di ffi -cult and prone to large uncertainties. Individual B-type stars lackage-related chromospheric indicators and can essentially only bedated using isochronal fitting. However, B-type stars also exhibithigh degrees of multiplicity and rapid rotation, both of whichneed to be accurately determined; otherwise ambiguities in theisochronal analysis are introduced.In BEAST, we are greatly helped by the fact that all ofthe target stars (except for the outliers discussed in Sect. 2.2)are high-probability members of Sco-Cen, which has a well-determined statistical age. The US, UCL, and LCC sub-regionshave estimated ages of 10 ±
7, 16 ±
7, and 15 ± ff erentialsof a factor ∼ ff ectively provides an age estimate that is Article number, page 4 of 22. Janson et al.: BEAST survey based on isochronal dating of suitable Sco-Cen stars that are cor-related particularly well with the spatial position of the target.Another, yet more refined approach is to determine an agebased on isochronal dating of Sco-Cen stars that are particularlyclose to the target star in phase space – that is both in positionand velocity. We executed a dedicated such procedure for theBEAST sample, which is described in the following.
As the map from Pecaut & Mamajek (2016) demonstrates, thefull Sco-Cen region spans a range of ages, consistent with beingthe result of a complex series of smaller star formation eventsspanning several million years. Coeval stars within the grouphave correlated positions as a result, which can be utilized whenextracting ages based on location in the map. However, stars aris-ing from di ff erent formation events gradually intermix over time,so the sky-projected location of a Sco-Cen star is in principle anincomplete proxy for its age. This implies that improvements inage precision and / or accuracy could be achieved by accountingnot only for correlations in physical space, but also for correla-tions in velocity space.Thus, in this section we examine each BEAST target forthe purpose of trying to identify nearby stars with very simi-lar 2D-projected space motions. We refer to such stars as co-moving stars (CMS). A CMS could be a wide, physically boundcompanion to the target star, but it does not need to be; we donot attempt to distinguish which CMS are physical compan-ions and which are not. Rather, we see the identified groups ofstars as co-moving streams originating from a common extendedstar formation event, and which will likely form the seeds fordistinct young moving groups (e.g. Zuckerman & Song 2004;Torres et al. 2008) at older ages, once Sco-Cen as a whole hasdispersed beyond straightforward recognition. Gaia is of criti-cal importance for this analysis, since it provides proper motionsfor a large number of Sun-like and lower-mass members of Sco-Cen that were previously impossible to identify, which can nowbe subjected to CMS analysis. In principle, usage of RVs wouldfurther enhance the analysis, since it would add a dimension ofmovement and also enable convergence tracing. However, sincethe vast majority (90%) of the CMS candidates in our analysisdo not have RV measurements yet, such an analysis cannot beperformed at the present time.We start our CMS analysis by transforming the
Gaia propermotions and parallaxes into sky-projected velocities v α and v δ and heliocentric distances d . The physical separation ξ betweena candidate CMS with distance d i and a target star with distance d can then be calculated through ξ = q d i + d − d i d cos ( ∆ θ ) , (1)where ∆ θ is the angular separation between the two objectson the sky. Likewise, we can calculate a 2D-projected di ff erentialvelocity ∆ v between the candidate CMS (velocity v i ) and target(velocity v ) as ∆ v = q ( v α, i − v α, ) + ( v δ, i − v δ, ) . (2)Similarly as in, for example Röser et al. (2018) andMeingast et al. (2019), we set a threshold of ∆ v < . / s fordistinguishing between CMS and non-CMS candidates. Since 1 km / s corresponds approximately to 1 pc / Myr and Sco-Cen is ap-proximately 15 Myr old, it follows that only stars within ∼
20 pcof a given target star can be a viable CMS, so we can set as anadditional threshold that ξ <
20 pc. Finally, to avoid relying onpotential CMS stars with large uncertainties, we exclude stars forwhich either (cid:12)(cid:12)(cid:12)(cid:12) v α − . σ v α (cid:12)(cid:12)(cid:12)(cid:12) < (cid:12)(cid:12)(cid:12)(cid:12) v δ − . σ v δ (cid:12)(cid:12)(cid:12)(cid:12) <
2, that is stars that cannotbe distinguished from the threshold at more than 2 σ confidence.Likewise, we exclude stars with π ln · σ π > .
07, that is stars forwhich the uncertainty in the distance would necessitate absolutemagnitude errors greater than 0.07 mag. The latter criterion ismotivated by the fact that we need good precision photometryfor the isochronal analysis.We show the distribution of targets and CMS in Fig. 4. In-cidentally, their placements also cleanly outline the three sub-regions of Sco-Cen. A specific example of a target (HIP 60009)and its associated CMS candidates is shown in Fig. 5.
Fig. 4.
Sky distribution of all the targets in BEAST (red symbols) andtheir identified CMS candidates (blue symbols).
Fig. 5.
Sky placement and proper motion vectors of HIP 60009 (in red)and its associated CMS candidates (in black). The length of the arrowsis proportional to the speed of the proper motion, where the size of thearrow in the legend denotes 20 mas / yr. Article number, page 5 of 22 & Aproofs: manuscript no. main
For each candidate CMS, we collected photometry from
Gaia (Brown et al. 2018) and (Skrutskie et al. 2006). We cre-ated four independent sets F of photometric pairs: { G , J } , { G , H } , { G , K } , and { Gbp , Grp } . We corrected for reddening us-ing the STILISM (Capitanio et al. 2017) 3D maps to acquire an E ( B − V ) value for each individual CMS, adapting the redden-ing to each relevant photometric band using conversions fromWang & Chen (2019). Each photometric pair can be comparedwith theoretical isochrones in order to constrain the age andmass of a given CMS candidate. The BT-SETTL (Allard 2014)isochrones are used for this purpose. The data points are chosenas pairs since two parameters are required to resolve the ambigu-ity between mass and age, and the pairs are chosen on the basisof combining a large wavelength span with small photometricerrors. In the k -th photometric band, we can now relate the mea-sured flux F k to model fluxes F k , model . Given that our model canbe formulated in a compact form as a 3D matrix, whose elements { a i jk } represent the fluxes corresponding to the i -th mass, the j -thage, and the k -th filter, we can construct a 2D distance matrixwhose elements δ i j are defined as δ i j = vtX k ∈F F k − F i jk , th ∆ F k ! , (3)where ∆ F k is the photometric error. In this context, δ min = min δ i j represents the best approximation to the measured fluxes,yielding the most probable mass and age of the CMS. The upperthreshold for constituting an acceptable fit was set at δ min = k -th band ifits flux error was above 10% (0.1 mag).With an average number of ∼
36 suitable CMS candidates pertarget (see Table 4), it is in principle possible to get very preciseage estimates for the targets by averaging the isochronal agesof their respective CMS samples. However, several considera-tions need to be taken into account for acquiring maximally ac-curate estimates. As a first step, we note that the CMS matchingto the targets is based on a sharp cut-o ff of CMS versus non-CMScandidates, but in reality there is a gradient where the probabil-ity gradually decreases outward from the closest, most probableCMS companions. We account for this by calculating a weightedmean of the CMS ages, imposing weights w i defined as w i = ∆ v max ξ max ξ ! + ( ∆ v α ) + ( ∆ v δ ) − / . (4)In this equation ∆ v max = . / s and ξ max =
20 pc arethe maximum allowed relative velocity and separation betweena target and CMS as described in the previous section. Since theisochrones are sampled logarithmically in age, we calculate aweighted mean m on the logarithmic ages rather than the linearages (equivalent to a geometric mean) as follows: m = X i ln ( t i ) ˜ w i , (5)where t i are the individual ages and ˜ w i the normalizedweights (i.e. normalized to a sum of 1). The corresponding un-certainty s then becomes s = sX i ˜ w i ( ln ( t i ) − m ) · nn − X i ˜ w i , (6) where nn − is the Bessel correction factor for CMS samplesizes of n >
1. Since we perform the calculations on four di ff er-ent filter pairs, we get four di ff erent means and uncertainties m j and s j . This can be translated back into linear quantities through µ j = exp ( m j ) and σ j = µ j (exp ( s j ) − s j for the weights; w j = / s j = / ( ln ( σ j /µ j ) + ffi ciently accurate. However, thereare known biases in the sample that skew the isochronal analysisunless properly accounted for.In particular, there are two primary aspects of the target sam-ple that can potentially impose biases in the CMS-determinedages: the age-mass bias and the multiplicity bias. The formeris based on the observed fact that low-mass stars in youngstellar regions such as Sco-Cen feature systematically youngerisochronal age estimates than Sun-like and higher-mass stars(e.g. Pecaut & Mamajek 2016). The discrepancy gets increas-ingly pronounced towards later spectral types across the M-typesequence. A priori, this could hypothetically be interpreted as atrue mass-age gradient in the systems, with low-mass stars form-ing later than higher-mass stars. However, the discrepancy per-sists even within physically bound stellar systems that shouldbe expected to be coeval (Asensio-Torres et al. 2019). This im-plies that young low-mass stars have a systematic bias in theirisochronal age estimates, which theoretically can be understoodas an incomplete treatment in conventional models of the in-fluence of magnetic fields on the size and temperature on thestars (Somers & Pinsonneault 2015; Feiden 2016). Since manyof our CMS are low-mass stars (approximately reflecting themass function in Sco-Cen), it is important to calibrate our proce-dure to account for this e ff ect. Ideally, this would be done us-ing models that perfectly account for the magnetic fields andthe resulting chromospheric structure of low-mass stars. How-ever, complete sets of such models do not exist yet. While thereare isochronal models that do attempt to account for magnetice ff ects in late-type stars (Somers et al. 2020), these models takethe spot coverage of the star as an additional free parameter. Thisfree parameter is relatively coarsely sampled in the isochrones,fundamentally unknown for the CMS candidates in our sample,and probably itself dependent on the age of the star. For thesereasons, we opt to instead address the bias imposed by conven-tional isochronal models (in this case represented by BT-SETTL)in an empirical way.A straightforward way to avoid any mass-age bias relativeto the Pecaut & Mamajek (2016) map estimates would be to usethe same mass range for our CMS as was used for constructingthe map (i.e. ∼ M sun and higher). However, imposing such ahigh-mass threshold would remove most of the CMS and preventany age analysis to be performed on targets with relatively fewCMS associated with them. Instead, we calculated a statisticalcorrection factor β based on the targets with the most numeroussamples of CMS ( n > β values for such targetswere calculated as the quotient between an age estimate t c basedonly on > M sun CMS versus an age estimate t based on allCMS for that target. Averaged over all n >
50 targets, we getˆ β = . ± .
03. This factor was then uniformly applied to all(non-thresholded) age estimates in the sample, correcting for theage-mass bias.The multiplicity bias in our sample refers to the fact that anyunresolved multiplicity makes a CMS look artificially brighter,which in turn leads to a bias in its age estimation. We alreadydiscussed this as an important limitation for isochronal dat-
Article number, page 6 of 22. Janson et al.: BEAST survey ing of the B-stars themselves. For the lower-mass CMS candi-dates, the multiplicity fraction is lower than for high-mass stars(Duchêne & Kraus 2013); therefore the e ff ect is smaller, but stillpresent if not accounted for. Since a typical CMS is in the M-type spectral range, we can expect a total multiplicity fraction of ∼ ∼
30% to havea magnitude di ff erence between the primary and secondary of < ff erences, it follows that wecan expect a small fraction of the systems to be heavily influ-enced by the multiplicity bias, while the majority of the sampleare negligibly a ff ected or una ff ected by this e ff ect. On this topic,it can also be noted that a small but non-negligible fraction ofthe CMS can be expected to host protoplanetary discs, whichare more common among low-mass stars than for higher-massstars at Sco-Cen ages (Carpenter et al. 2006; Luhman & Esplin2020). The presence of such a disc sometimes causes a CMS toappear fainter and thus older in an isochronal analysis, producingoutliers in the opposite direction from the multiplicity bias.Owing to the outlier nature of both of these biases, they couldbe easily accounted for by taking a median (rather than a mean)of the CMS ages to get a representative age estimate for the cor-responding target star. However, this has the drawback that itwould not be possible to use the weighted mean scheme out-lined above, which was developed to increase the accuracy ofthe ages. Therefore, we opt for an intermediate solution basedon percentile rejection: For each CMS distribution, we reject thetop 10% and bottom 10% estimated ages (except if n ≤
6) andcalculated a weight mean based on the remaining values to es-tablish the best-fit age of each target star. This retains the robustproperties of the median in rejecting outlier ages, whilst simul-taneously keeping the weighting scheme intact.The ages resulting from this analysis including bias mitiga-tions are shown in Table 3. The average age across the sample forthis method is 16.5 ± ± >
95% confidence.
In Table 3, we summarize age estimations for the tar-get sample using three di ff erent methods: (1) Average sub-group age (SG method); (2) age based on interpolation ofthe Pecaut & Mamajek (2016) map (MAP method); and (3)isochronal dating of co-moving stars (CMS method). The di ff er-ent methods complement each other rather well since, for exam-ple not all targets have identified co-moving stars associated withthem, and not all stars are covered by the Pecaut & Mamajek(2016) map. In cases for which all three are available, we planto use CMS as the first priority, MAP as second priority, andSG as third priority for statistical purposes. For individual tar-gets with well-determined multiplicity and rotational properties,direct isochronal dating is of course another option. Fig. 6.
Cumulative distributions for the target ages as estimated throughour CMS analysis (black line) and through the Pecaut & Mamajek(2016) age map (red line). The two methods show a high degree ofconsistency. Also shown as a blue line is the distribution of ages froma direct isochronal analysis for some of the B-type stars in the sam-ple, as determined in Pecaut & Mamajek (2016) based on models fromEkström et al. (2012). This latter method is less precise for a typicalstar in this sample relative to the other methods, and thus not used inthis paper.
While the precise stellar mass in a given system is not criticalto determine the properties of planets in the system (unlike thesystem age), it is an important parameter for statistical purposes;this is the case both for the purpose of determining planet occur-rence as a function of stellar mass and determining mass ratiosbetween detected companions and their hosts. A complicatingfactor in determining stellar mass in a B-type sample is that themultiplicity fraction for such systems is very high; the total mul-tiplicity is 50–60% or higher according to the literature reviewof Duchêne & Kraus (2013). If a stellar multiple is unresolvedwith an unknown mass / flux ratio of the two (or more) compo-nents, then its component masses cannot be unambiguously de-rived from isochronal analysis. However, if a multiple systemcan be distinguished and monitored over a su ffi cient baseline,it can become possible to determine dynamical masses withoutany model uncertainty.While many of the Sco-Cen B-stars are multiple (see Sec.2.5), most of these have no dynamical mass determinations asof yet. Hence, to estimate stellar masses in our sample, we usedthe spectral type (SpT) of each system as a proxy for the massof its primary star. For this purpose, we used exactly the samerelation that has already been used for early-type stars in Sco-Cen in Lafrenière et al. (2014). The resulting mass estimationsare summarized in Table 5 and in Fig. 7.To test the accuracy of these estimations for the BEASTsample, we cross-checked them against the three systems ofthe sample that are both known eclipsing binaries and double-lined spectroscopic binaries, such that individual masses canbe determined in an entirely model-free manner: HIP 74950(Budding et al. 2015), HIP 78168 (David et al. 2019), and HIP82514 (Budding et al. 2015). The primary masses as determineddynamically are 4.16, 5.58, and 8.3 M sun respectively, while thecorresponding SpT-inferred primary masses are 3.3, 5.9, and 9.0 M sun . For all the tested cases, the estimated masses are thus Article number, page 7 of 22 & Aproofs: manuscript no. main within 6–21% of the measured value without any noticeable sys-tematic o ff sets, so we consider this to be su ffi cient precision forthe purpose of primary mass estimation.We note that for most close binaries, for which any imagedplanetary companion would be circumbinary, very little is knownabout the mass of the secondary star. This can lead to large un-certainties if the mass ratio is calculated with respect to the to-tal central mass, as seen for HIP 79098 (AB)b in Janson et al.(2019). On a longer timescale, it is therefore highly desirable tobetter characterize the orbits of the close binaries in our sample.Radial velocity, Gaia DR >
2, interferometry, and high-resolutionimaging could all be useful techniques for this purpose. N u m be r o f t a r ge t s Fig. 7.
Histogram of estimated stellar masses for the BEAST sample.
To assess multiplicity over a wide range of semi-major axes, acombination of several di ff erent methods must be applied. Forclose-in binaries (periods of days to years), RV is the most ap-plicable technique, although the broad spectral lines of B-typestars limit the achievable precision and make it more di ffi cult toidentify double-lined binaries. While there is no published co-herent survey for RV multiplicity among B-stars across all ofSco-Cen, many of the targets have been observed over at least afew epochs with RV, sometimes stretching back several decades,owing to their high visual brightnesses. We scanned the litera-ture for indications of RV binarity both in surveys and individualstudies and found 22 such cases. As already mentioned in Sect.2.4, three of these binaries simultaneously exhibit double linesand are eclipsing; in these cases, detailed system parameters canbe derived. We also identified two other eclipsing binaries in thesample from the literature search: HIP 67464 has an eclipse pe-riodicity of 2.6 days (Dubath et al. 2011), and is also listed asa spectroscopic binary in the SB9 catalogue of spectroscopic bi-nary orbits (Pourbaix et al. 2004) with the same period, althoughseemingly only as an SB1. HIP 65112 is a system for whicheclipses have been reported (Malkov et al. 2006; Dubath et al.2011), but for which no reported instances of RV binarity existin the literature. Essentially all of the close RV binaries o ff er theopportunity to determine unique component masses and similarconstraints over short timescales in principle, but more data thanis currently available would be required, for example from Gaia astrometry or interferometric imaging. Intermediate-separation binaries (a few AU to hundreds ofAU) are primarily detected with interferometry in the closer-in cases and AO imaging in the wider cases. In both circum-stances, the WDS catalogue lists companions from the literaturewith a high degree of completeness. We removed known com-panions in the 0.1–6 ′′ range from the BEAST sample as dis-cussed in Sect. 2.1. However, systems that only contain com-panions with smaller separations than 0.1 ′′ were kept in thesample. These have primarily been detected through interfer-ometry (Rizzuto et al. 2013), but in some cases also with AO(e.g. Shatsky & Tokovinin 2002). The literature AO surveys alsoyielded some confirmed physical companions outside of 6 ′′ .Since previous surveys were not fully completed across the Sco-Cen region, we also found a number of new binary companionsin our BEAST observations in the 0.1–6 ′′ region where theywould have been removed from the sample if they had beenknown beforehand. These new binaries are discussed in moredetail in Sect. 5.3.Wide binaries at separations in the thousands of AU canonly be detected in wide-field astrometric surveys such as with Gaia , by identifying objects that share a common proper mo-tion (CPM) with the target star. As we have seen in Sect. 2.3however, there are many stars in Sco-Cen that share very simi-lar proper motions with our individual target without necessarilybeing bound. In this paper we do not attempt to distinguish be-tween wide binaries and other CMS stars. But in Table 4, whichsummarizes the (known) multiplicity properties of the sample,we also make a note of the number of CMS stars associated witheach target from our isochronal analysis for general reference.The search volume for that analysis had a 20 pc radius, so mostCMS stars are not expected to be physically bound.In principle, proper motion accelerations over time for in-dividual targets between
Hipparcos and
Gaia , for example,could also be used to identify and characterize companions(e.g. Calissendor ff & Janson 2018; Brandt 2018; Kervella et al.2019). However, this requires a very high precision in gen-eral, so given the large excess uncertainties discussed in Sect.2.2, the DR2 data are not yet su ffi cient for a robust analysis ofthe BEAST sample; subsequent Gaia releases should be bettersuited for that purpose.
Beyond their direct connection to planet formation and evolu-tion, the presence and properties of discs among target stars in di-rect imaging studies are relevant for several purposes: The pres-ence of bright discs may be correlated with the presence of widemassive planets (Meshkat et al. 2017), and if there are signs ofgaps in the discs, that may give prior information about wherethe planets could be located (e.g. Apai et al. 2008; Janson et al.2013a). Furthermore, the disc may be detected in high-contrastimaging if it is su ffi ciently bright, allowing for its detailed mor-phology to be mapped; this can provide further informationabout planets in the system (e.g. Chiang et al. 2009; Dong et al.2018) and about the disc itself (Boccaletti et al. 2015; Milli et al.2017).For our sample, there are two main ways to identify discs:The dust in the disc gives rise to some degree of infrared excess,and if there is gas in the disc, it gives rise to emission lines, gen-erally leading to a Be-type spectral classification for the star. Wescanned the literature for both kinds of indicators. Infrared ex-cess is usually inferred from Spitzer or WISE data, or both (e.g.Luhman & Mamajek 2012; Rizzuto et al. 2012). We excludedcases in which the inferred properties from the spectral energy
Article number, page 8 of 22. Janson et al.: BEAST survey distribution (SED) fitting are odd (e.g. unrealistically high ex-tinction) and there is only one data point to indicate the excess.The SED fitting is relatively di ffi cult for B-type stars, partly be-cause the star itself is very bright even at infrared wavelengthsso that a relatively large excess is required to be detectable andpartly because the multiplicity rate is high, which complicatesthe fitting. Nonetheless, we inferred a reasonable and significantexcess for 21 of the BEAST targets, which is 25% of the sample.Be-type stars are generally classified as such in SIMBAD,but since these classifications sometimes date back to sourcesfor which the underlying data is not necessarily presented, we re-quire that a star shows significant emission in at least one of theepochs presented in Arcos et al. (2017) to count the star as a disccandidate in our sample. Both the excesses and Be-type classi-fications are shown in Table 5. While line emission seems lessfrequent (six instances, 7% of the sample) than excess, the twoproperties correlate well, in the sense that stars with identifiedemission almost always occur in systems with identified excess.This fact supports the notion that both properties are probablygood indicators for discs in the systems. In both cases, the per-centages given should be seen as merely indicative and shouldbe regarded as lower limits, because the literature surveyed isnot necessarily complete and our rejection process was conser-vative in order to include only reasonably secure cases.
3. Observations
The BEAST survey started running during mid-2018, and hasconsistently used SPHERE (Beuzit et al. 2019) for its observa-tions. At the time of writing, most first-epoch data (81%) havebeen acquired, while almost all of the second epoch observationsare allocated but remain to be executed. All first-epoch observations are acquired in the IRDIFS-EXTmode, in which the
Y JH -band wavelength range is covered bythe integral field spectroscopy (IFS) arm of the instrument, andthe K -band range is covered by the IRDIS arm. The IFS armprovides integral field spectroscopy in a ∼ . × . ∼ ×
11 arcsec FOV.With its coverage in the K -band range through the K K ∼
100 mas).The integration times are based on a statistical prediction of am-bient conditions; during the actual observations, the conditions The VLT has been closed down during much of 2020 as a result ofthe COVID-19 pandemic. can be such that some saturation still occurs at the coronagraphedge. This is considered an acceptable trade-o ff .Before and after each main ADI sequence, a non-saturatedimage of the primary star (using neutral density filters) is ac-quired, and a “wa ffl e” image is also acquired, in which the staris behind the mask but the wa ffl e mode of the deformable mirroris turned on, such that ghost images appear at specific locationsthat can be used for both astrometric and (in principle) photo-metric calibration. A set of sky frames is also acquired, whichis particularly important in IRDIFS-EXT observations since theIRDIS arm operates in the K -band range where the thermal back-ground is relatively high. Dithering of the IRDIS detector is usedto account for detector e ff ects such as bad pixels.Since Sco-Cen as a whole is relatively close to the Galacticplane, background sources are commonplace, therefore follow-up observations need to be executed in the majority of casesto distinguish background stars from real physical companions.Most follow-up observations will be executed in an identical wayas the first-epoch observations, with occasional exceptions suchas if a disc candidate is detected, in which case the follow-upmay be performed at a shorter wavelength and with polarimet-ric di ff erential imaging (PDI; see e.g. Hashimoto et al. 2011;Schmid et al. 2018) in addition to ADI. Since a primary goalwith the second epoch is generally to test for CPM, and sinceSco-Cen members typically have a relatively modest proper mo-tion of 20–30 mas / yr, we aim for all second epoch observationsto be acquired with at least a one-year baseline w.r.t. the firstepoch. Further follow-up, such as detailed spectroscopic charac-terization, can then be done in a third epoch once CPM is es-tablished; these follow ups are usually outside of the large pro-gramme, in which the main part of the programme is run. If thecompanion happens to reside in the IFS FOV, then spectroscopiccharacterization is possible even in the first and second epochsof data. In this paper we focus primarily on the first epoch ob-servations.An observing log of the first epoch observations is shownin Table 1. In ESO’s service mode, the observations are asso-ciated with a set of constraints on the ambient conditions thatneed to be fulfilled for the output data to be scientifically useful.Observations are typically only executed if all the constraintsare fulfilled; if conditions deteriorate significantly during an ob-servation, it is discarded and re-executed at a later time. In ourcase, the most central constraint is that atmospheric conditionsshould have a grade of “good” or better in the ESO classifica-tion scheme, which corresponds to a seeing better than 0.9 ′′ anda coherence time better than 4 ms , apart from a timing con-straint requiring that the observations must take place arounda meridian transit. In the ESO grading system, “A” means thatall the constraints were fulfilled, while “B” means only smalldeviations from the requirements. Observations graded “C” orlower are considered not scientifically useful and scheduled forre-execution. In total, 67 first-epoch observations with an “A” or“B” grading have been acquired to date, which are included withtheir respective grading in the observing log.
4. Data reduction and analysis
The fundamental data reduction process for the survey is per-formed in a primarily automatic and streamlined manner, usingthe SPHERE Data Center (DC; see Delorme et al. 2017). Basicdata reduction steps include dark and flat corrections; pixel scale, For the earliest observations, the seeing criterion was 0.8 ′′ , but werelaxed it slightly from 2019 onwards to allow for easier scheduling.Article number, page 9 of 22 & Aproofs: manuscript no. main true north corrections, and distortion corrections from observa-tions of stellar clusters; image cube creation; and wavelengthcalibration. The true north angle is typically -1.8 deg and is de-termined on a run-by-run basis with a precision of ∼ / pixel forIRDIS and 7.46 mas / pixel for IFS and these quantities are deter-mined each run with a precision of 0.01–0.02 mas / pixel. Afterthe calibrated image cubes are produced, the high-contrast pro-cessing begins, which uses the SpeCal package (Galicher et al.2018) within the context of DC as well. SpeCal allows for a widerange of possible options in terms of optimization algorithms andcorresponding parameter settings.As part of our standard reduction procedure, we use threeseparate algorithms for IRDIS and three for IFS for each tar-get. For IRDIS, the three approaches are as follows: (1) A rota-tion and collapse of the time series without any ADI subtraction,which allows for a very quick overview of the field and the dataquality. (2) A classical ADI approach with a median across thetime series representing the point spread function (PSF) model,which yields a conservative reduction that has benefits whensearching for faint extended emission (debris discs) that can eas-ily get partially subtracted in a more aggressive scheme. (3) ATemplate Localized Combination of Images (TLOCI) approach,which is optimized for finding point sources in the images. TheIFS procedure is similar, but because of its normally entirelycontrast-limited FOV, a non-ADI approach is not very mean-ingful. Thus the three approaches are as follows: (1) a classi-cal ADI approach; (2) a TLOCI approach; and (3) a Karhunen-Loève Image Projection (KLIP, see Soummer et al. 2012; Pueyo2016) Principal Component Analysis (PCA)-based approach asan alternative to the TLOCI algorithm. Beyond this standard-ized procedure, several e ff orts are in progress to analyse theBEAST data with alternative reduction schemes, including Pyn-Point (Stolker et al. 2019), ANDROMEDA (Cantalloube et al.2015), and TRAP (Samland et al. 2020), but this paper focusseson the standardized procedure.After the final reduced images are produced, they are visu-ally inspected and convincing point sources at ≥ σ are identi-fied. All of these companion candidates (CCs) are analysed withthe characterization module of SpeCal, in a second step of DCprocessing. This yields K K K − K
5. Results and discussion
We show the 5 σ contrast performance as a function of sepa-ration for IRDIS in Fig. 8 and for IFS in Fig. 9. For exam-ple, at 0.4 ′′ , the median contrast across the observed sample is9 × − (12.6 mag) for IRDIS and 4 × − (13.5 mag) for IFS.This is deeper thanthe GPIES (Nielsen et al. 2019) and SHINE(Langlois et al. 2020) surveys, for instance. The contrast perfor- mance results from a combination of the excellent performanceof SPHERE, the relatively good (but not exceptional) averageobserving conditions, and the high average brightness of thestars in the sample, allowing for a good wavefront sensing per-formance in each case. This contrast resulted from a baselineSpeCal reduction optimized for uniformity across the sample.It would most likely be possible to enhance the contrast furtherstill on a star-by-star basis through alternative high-contrast al-gorithms (e.g. Samland et al. 2020) or through adjusting the pa-rameters in the SpeCal algorithm. As mentioned in Sect. 4, ex-periments to this end are ongoing in parallel to the survey-scalereduction e ff orts. The BEAST sample is nearly ideal for such ex-periments owing to the widely extended contrast-limited regimesof the bright host stars. C on t r a s t ( m ag ) IRDIS TLOCI contrast
Fig. 8.
Contrast curves for IRDIS with the TLOCI reduction. The K K C on t r a s t ( m ag ) IFS TLOCI contrast
Fig. 9.
Contrast curves for IFS with the TLOCI reduction. The collapsedimages across the accessible wavelength range is used. Grey lines: Con-trast curves for individual targets. Blue thick line: Median contrast curvefor the sample.Article number, page 10 of 22. Janson et al.: BEAST survey
The observations presented in this work contain a total of 708faint candidates inside the IRDIS FOV of the 67 observed tar-gets. As expected from the varying stellar backgrounds, thesecandidates are unevenly distributed among the targets, of which11 have empty fields without any candidates; the most crowdedfield has 117 candidates. All of the candidates will requirefollow-up to test for CPM in a second epoch before it can becategorically established whether they are physical companionsor background stars. This was expected from the outset and isa part of the allocated BEAST observational programme. Sincethe dual-band imaging mode of IRDIS is used, K K K − K K K K
1. Nonetheless, particularly high-merit candidates may poten-tially be identified through unusually blue (implying molecularabsorption) or red (implying clouds) colours. A collective CMDfor all candidates with measured K − K Fig. 10.
Colour-magnitude diagram in M K vs. K − K ff erent colours as shown inthe legend. Previously discovered low-mass substellar companions thathave been observed in the K K When a point source around a target star is bright enough to bea stellar companion, it has an extremely low false positive prob-ability in general, so it can already be regarded as a probablecompanion from the first epoch. We thus present the astromet- ric properties of the companions that have been detected in thestellar mass / brightness regime in Table 6 . Their statistical prop-erties will be examined in more detail in Squicciarini et al. (inprep). In total, we identified ten stellar companions, of which sixhad not previously been identified in the scientific literature. Webriefly discuss each individual case below. – HIP 50847: The presence of a stellar companion to HIP50847 was tentatively hinted at in an observation taken fora multiplicity study of Sco-Cen presented in Janson et al.(2013b). However, the corresponding data set was taken atambient conditions far inferior to the programme require-ments, and therefore the reduced image was of so poor qual-ity that no actual inference could be made about the verac-ity of such a companion. The BEAST data set is the first inwhich a companion can be robustly inferred (see Fig. 11).Aside from the visual binarity, HIP 50847 is also known asa double-lined spectroscopic binary with a period of only 15days (Quiroga et al. 2010), meaning that there are three con-firmed stellar components in the system.
Fig. 11.
Non-ADI IRDIS image of HIP 50847, with the primary com-ponent denoted “A” and the new companion at intermediate separationdenoted “B”. – HIP 59173:
This star was previously observed withAO imaging in a survey by Shatsky & Tokovinin (2002).The companion detected in BEAST is not reported inShatsky & Tokovinin (2002), which is probably because thecompanion is a low-mass star. Its contrast relative to the pri-mary star may have been challenging for the first-generationAO system ADONIS used in that survey. In the literature,HIP 59173 is noted as a double-lined spectroscopic binary(Chini et al. 2012); no period is reported, but since the linescan evidently be kinematically distinguished in the RV data,the spectroscopic binarity must refer to a much closer-incompanion than the BEAST-detected companion at ∼ – HIP 60009: Very similarly to HIP 59173, HIP 60009 wasobserved with ADONIS in the Shatsky & Tokovinin (2002)survey, but the companion detected in BEAST was not re-ported in their study. In this case, presumably the small sep-aration of ∼
150 mas of the companion would have made it With the exception of HIP 76600, see individual note.Article number, page 11 of 22 & Aproofs: manuscript no. main di ffi cult to detect with ADONIS. Also very similarly to HIP59173, HIP 60009 has been noted as a double-lined spec-troscopic binary in the literature (Chini et al. 2012), so as awhole the system is probably at least triple. An image of thestellar system is shown in Fig. 12. Fig. 12.
Non-ADI IRDIS image of HIP 60009, showing the primary(“A”) component as well as the newly discovered close companion(“B”). – HIP 62434:
According to our archival search, HIP 62434has never been observed with AO or similar facilities, andthus its intermediate-separation multiplicity has been largelyunprobed. As we note in Table 4, this object has been subjectto RV studies in the past (Pourbaix et al. 2004), which haverevealed a spectroscopic companion with a period of 1828days. Our detected companion from BEAST images has aprojected separation of ∼
357 AU and is therefore certainlydistinct from the RV companion. Hence, the HIP 62434 sys-tem is at least a stellar triple. – HIP 63005:
Similar to HIP 59173 and HIP 60009, HIP63005 was observed in the Shatsky & Tokovinin (2002) AOsurvey, but the companion detected in the BEAST data wasnot reported there. The BEAST companion resides at 274mas separation, which would have made it di ffi cult to seewith ADONIS. HIP 63005 is in a very wide binary pairwith the B-star HIP 63003, which is also part of the survey;otherwise there are no additional stellar companions knownaround either star. – HIP 63945: The companion discovered in BEAST waspreviously known, as was reported in Shatsky & Tokovinin(2002). Normally, such an identified intermediate-separationstellar binary would have been discarded in the BEASTmaster list construction, but in this case it had been over-looked owing to the relatively complex multiplicity listingfor this particular target in the WDS database. One of thereasons for this is that the system additionally has an in-ner interferometrically detected companion (Rizzuto et al.2013), which was close-in enough to be allowed within theBEAST selection criteria. HIP 63945 also has a noted single-line spectroscopic binary companion of undetermined period(Chini et al. 2012), which may or may not be identical to theinterferometric companion. Like many of the other visual bi-naries reported in this paper, the system is thus a high-ordermultiple with at least three components. – HIP 73624:
HIP 73624 was originally included in theBEAST sample, but when a bright stellar companion wasdiscovered in the IRDIS FOV during a short aborted run, itwas deemed a suboptimal target and removed from the sam-ple. Hence, this target is not listed in other BEAST tables, butfor completeness we include the location of the newfoundstellar companion in Table 6. – HIP 74100:
This target has a companion noted inShatsky & Tokovinin (2002), but it was kept in the originalsample because the quoted separation in the WDS cataloguewas well within the range of where the companion wouldhave impacted the observational performance of SPHERE.Since a companion was detected at a 550 mas separation inBEAST, we revisited the literature. We note that the sepa-ration, 6 mas, quoted in Shatsky & Tokovinin (2002) mustpresumably be a typographical error because the observa-tions were performed with ADONIS whose angular reso-lution was far worse than ∼ – HIP 76600: There is an interferometric companion to thistarget noted in Rizzuto et al. (2013), which is marginally de-tected in BEAST images, however not to the extent that reli-able astrometric properties can be determined. We thereforeomit this target from table 6. There is also a spectroscopiccompanion with a period of only 3.3 days (Pourbaix et al.2004), which is therefore a distinct third stellar componentin the system.
Fig. 13.
Non-ADI IRDIS image of HIP 81972. The component nota-tion follows the notation from previous epoch imaging of the system(Shatsky & Tokovinin 2002). We find that “C” is a background star,while “E” is a real physical companion. – HIP 81792:
HIP 81792 resides in an unusually crowded re-gion of the sky, and thus has a much higher spatial den-sity of bright contaminants relative to most other targets.In the IRDIS FOV, there are two candidates that are bothbright enough to be stellar even if co-moving; HIP 81972is the only target in the observed sample for which this isthe case. Because of these special circumstances, we can-not necessarily infer that they are probable companions de-spite their high brightness. Fortunately, HIP 81972 was ob-served by Shatsky & Tokovinin (2002) and both candidates
Article number, page 12 of 22. Janson et al.: BEAST survey are reported as visible in their data, so we have the opportu-nity to test whether they are physical companions through aCPM test. We designate the candidates “C” and “E’,’ respec-tively, following the notation in WDS. The CPM test for “C”is shown in Fig. 14, and the test for “E” in Fig. 15. Some-what counter-intuitively, the data show that the brighter andcloser-in candidate “C” is most likely a background contam-inant, while the slightly fainter and more distant HIP 81972E appears to be a genuine physical companion to HIP 81972A. Being identified as a likely background source, “C” is notincluded in Table 6; its separation and position angle in the2019 BEAST epoch are 2107.1 ± ± ffi ciently separatedfrom HIP 81972 A to have its distinct Gaia entry (Gaia DR25968351361342059264) and its proper motion and parallaxreported in
Gaia imply that it is a physically unrelated back-ground source. We show an image of the HIP 81972 systemin Fig. 13. y ( a r cs e c ) Fig. 14.
Proper motion test for the “C” candidate around HIP 81972.The black arrow denotes the direction of time for background motion.The motion of the candidate from the 2000 to the 2019 epoch is incon-sistent with CPM, but can be explained by a background star with asmall proper motion vector of its own.
6. Conclusions
The first large-scale, extreme-AO survey of planetary systemsaround B-type stars, BEAST will probe a nearly complete sam-ple of suitable B-type stars in the Sco-Cen region. This paper ex-plores and summarizes the properties of the BEAST stellar sam-ple. Important parameters for both interpreting individual sys-tems and statistical interpretations on a population basis includethe stellar masses, multiplicity properties, disc properties, andages. We place particular emphasis on the age aspect, where theSco-Cen region o ff ers some particularly appealing opportunitiesfor determining accurate ages. We develop a scheme for age de-termination of the B-type stars in the sample based on isochronal y ( a r cs e c ) Fig. 15.
Proper motion test for the “E” candidate around HIP 81972.The black arrow denotes the direction of time for background motion.For this case, the candidate is consistent with CPM, while simultane-ously being clearly distinct from the static background location. It is aprobable physical companion. dating of stars that can be closely kinematically matched to eachindividual target. This allows for a high precision in the age esti-mations, since an average of several co-moving stars can be ac-quired. The resulting ages are statistically consistent with otherage estimations from the literature, but as expected they are gen-erally more precise and most likely more accurate.First epoch observations were performed for 67 out of the 85targets in the survey, and these observations yielded several hun-dreds of candidate companions, which will need to be followedup in a second epoch to test for CPM. As in all direct imagingsurveys, the majority of the candidates are expected to be phys-ically unrelated background stars. However, if the frequency ofwide giant planets continues to increase with stellar mass in thesame way as the frequency increases in the M- to A-type stellarrange, then BEAST could potentially yield more new planet dis-coveries than any other direct imaging survey ever performed.The contrast limits provided in the survey are deeper on averagethan in comparable surveys, owing in large part to the perfor-mance of the SPHERE instrument, in particular for the brightstars that constitute the targets in the survey.We presented six new and four previously known binarycompanions detected within the survey. Since these detectionsare very bright, their false alarm probability is typically verylow, owing to the low probability of having a su ffi ciently brightstar in the background within a few arcseconds of the target. Anexception is the HIP 81972 system, which resides in front of anunusually crowded field, and as a result, does have a bright back-ground object in the IRDIS field of view. Previously reported inthe literature as HIP 81972 C, we tested the object for CPM usingliterature astrometry to create a 19-year baseline and found thatit is consistent with a near-static background star. On the otherhand, another star in the field, HIP 81972 E, is fully consistentwith being physically bound to the B-star primary. Acknowledgements.
M.J. gratefully acknowledges funding from the Knut andAlice Wallenberg Foundation. This study made use of the CDS services SIM-BAD and VizieR, and the SAO / NASA ADS service. R.G. and S.D. are supportedby the project PRIN-INAF 2016 The Cradle of Life - GENESIS-SKA (GeneralConditions in Early Planetary Systems for the rise of life with SKA). We also ac-knowledge support from INAF / Frontiera (Fostering high ResolutiON Technol-ogy and Innovation for Exoplanets and Research in Astrophysics) through the“Progetti Premiali” funding scheme of the Italian Ministry of Education, Uni-versity, and Research. Part of this research was carried out at the Jet Propulsion
Article number, page 13 of 22 & Aproofs: manuscript no. main
Laboratory, California Institute of Technology, under a contract with the NationalAeronautics and Space Administration (NASA). E.E.M. acknowledges supportfrom the Jet Propulsion Laboratory Exoplanetary Science Initiative and theNASA NExSS Program. This work has made use of the SPHERE Data Centre,jointly operated by OSUG / IPAG (Grenoble), PYTHEAS / LAM / CeSAM (Mar-seille), OCA / Lagrange (Nice), Observatoire de Paris / LESIA (Paris), and Obser-vatoire de Lyon / CRAL, and is supported by a grant from Labex OSUG@2020(Investissements d’avenir – ANR10 LABX56). G-DM acknowledges the sup-port of the DFG priority program SPP 1992 “Exploring the Diversity of Extra-solar Planets” (KU 2849 / References
Allard, F. 2014, Exploring the Formation and Evolution of Planetary Systems,299, 271Apai, D., Janson, M., Moro-Martín, A., et al. 2008, ApJ, 672, 1196Arcos, C., Jones, C. E., Sigut, T. A. A., et al. 2017, ApJ, 842, 48Asensio-Torres, R., Janson, M., Bonavita, M., et al. 2018, A&A, 619, A43Asensio-Torres, R., Currie, T., Janson, M., et al. 2019, A&A, 622, A42Beuzit, J.-L., Vigan, A., Mouillet, D., et al. 2019, A&A, 631, A155Boccaletti, A., Thalmann, C., Lagrange, A.-M., et al. 2015, Nature, 526, 230Bonavita, M., Desidera, S., Thalmann, C., et al. 2016, A&A, 593, A38Bowler, B. P., Liu, M. C., Shkolnik, E. L., et al. 2015, ApJS, 216, 7Brandt, T. D., McElwain, M. W., Turner, E. L., et al. 2014, ApJ, 794, 159Brandt, T. D. 2018, ApJS, 239, 31Brown, A.G.A., Vallenari, A. Prusti, T. et al. 2018, A&A, 616, 1Budding, E., Butland, R., & Blackford, M. 2015, MNRAS, 448, 3784Calissendor ff , P., & Janson, M. 2018, A&A, 615, A149Cantalloube, F., Mouillet, D., Mugnier, L. M., et al. 2015, A&A, 582, A89Capitanio, L., Lallement, R., Vergely, J. L., et al. 2017, A&A, 606, A65Carpenter, J. M., Mamajek, E. E., Hillenbrand, L. A., et al. 2006, ApJL, 651, L49Carpenter, J. M., Mamajek, E. E., Hillenbrand, L. A., et al. 2009, ApJ, 705, 1646Carson, J., Thalmann, C., Janson, M. et al. 2013, ApJ, 763, L32Chauvin, G., Vigan, A., Bonnefoy, M., et al. 2015, A&A, 573, A127Chauvin, G., Desidera, S., Lagrange, A.-M. et al. 2017, A&A, 605, L9Cheetham, A., Bonnefoy, M., Desidera, S. et al. 2018, A&A, 615, 160Chiang, E., Kite, E., Kalas, P., et al. 2009, ApJ, 693, 734Chini, R., Ho ff meister, V. H., Nasseri, A., et al. 2012, MNRAS, 424, 1925Cotten, T. H., & Song, I. 2016, ApJS, 225, 15Crepp, J. R. & Johnson, J. A. 2011, ApJ, 733, 126David, T. J., Hillenbrand, L. A., Gillen, E., et al. 2019, ApJ, 872, 161Delorme, P., Lagrange, A. M., Chauvin, G., et al. 2012, A&A, 539, A72Delorme, P., Meunier, N., Albert, D. et al. 2017, in SF2A-2017: Proceedings ofthe Annual meeting of the French Society of Astronomy and Astrophysics,ed. C. Reylè, P. Di Matteo, F. Herpin, E. Lagadec, A. Lancon, Z. Meliani, &F. Royer, 347-361Desidera, S., Chauvin, G., Bonavita, M. et al. 2020, A&A, submittedde Zeeuw, P.T., Hoogerwerf, R., de Bruijne, J.H.J., Brown, A.G.A., & Blaauw,A. 1999, AJ, 117, 354Dong, R., Najita, J. R., & Brittain, S. 2018, ApJ, 862, 103Dubath, P., Rimoldini, L., Süveges, M., et al. 2011, MNRAS, 414, 2602Duchêne, G. & Kraus, A. 2013, ARA&A, 51, 269Ekström, S., Georgy, C., Eggenberger, P., et al. 2012, A&A, 537, A146Eriksson, S. C., Asensio Torres, R., Janson, M., et al. 2020, A&A, 638, L6Feiden, G. A. 2016, A&A, 593, A99Gagne, J.,Lafreniere, D., Doyon, R. et al. 2014, ApJ, 783, 121Gagne, J., Mamajek, E., Malo, L. et al. 2018, ApJ, 856, 23Galicher, R., Boccaletti, A., Mesa, D. et al. 2018, A&A, 615, 92Ha ff ert, S. Y., Bohn, A. J., de Boer, J., et al. 2019, Nature Astronomy, 3, 749Hagelberg, J., Engler, N., Fontanive, C., et al. 2020, A&A, acceptedHashimoto, J., Tamura, M., Muto, T., et al. 2011, ApJL, 729, L17Helled, R., Bodenheimer, P., Podolak, M., et al. 2014, Protostars and Planets VI,643Janson, M., Bonavita, M., Klahr, H., Lafrenière, D., Jayawardhana, R., Zin-necker, H. 2011, ApJ, 736, 89Janson, M., Hormuth, F., Bergfors, C., et al. 2012, ApJ, 754, 44Janson, M., Brandt, T. D., Moro-Martín, A., et al. 2013, ApJ, 773, 73Janson, M., Lafrenière, D., Jayawardhana, R., et al. 2013, ApJ, 773, 170Janson, M., Bergfors, C., Brandner, W., et al. 2014, ApJ, 789, 102Janson, M., Asensio-Torres, R., André, D., et al. 2019, A&A, 626, A99Johnson, J. A., Aller, K. M., Howard, A. W., et al. 2010, PASP, 122, 905Kennedy, G. M., & Kenyon, S. J. 2008, ApJ, 673, 502 Keppler, M., Benisty, M., Müller, A. et al. 2018, A&A, 617, 44Kervella, P., Arenou, F., Mignard, F., et al. 2019, A&A, 623, A72Kuzuhara, M., Tamura, M., Kudo, T., et al. 2013, ApJ, 774, 11Lafrenière, D., Doyon, R., Marois, C., et al. 2007, ApJ, 670, 1367Lafrenière, D., Jayawardhana, R., Janson, M., Helling, C., Witte, S. &Hauschildt, P. 2011, ApJ, 730, 42Lafrenière, D., Jayawardhana, R., van Kerkwijk, M. H., et al. 2014, ApJ, 785, 47Lagrange, A.-M., Bonnefoy, M., Chauvin, G. et al. 2010, Science, 329, 57Langlois, M., et al. 2020, in prep.Lannier, J., Delorme, P., Lagrange, A. M., et al. 2016, A&A, 596, A83Liu, Q., Wang, T., & Jiang, P. 2014, AJ, 148, 3Lombart, M., Chauvin, G., Rojo, P., et al. 2020, A&A, 639, A54Luhman, K. L., & Mamajek, E. E. 2012, ApJ, 758, 31Luhman, K. L. & Esplin, T. L. 2020, AJ, 160, 44Macintosh, B., Graham, J.R., Barman, T. et al. 2015, Science, 350, 64Malkov, O. Y., Oblak, E., Snegireva, E. A., et al. 2006, A&A, 446, 785Marois, C., Lafrenière, D., Doyon, R., et al. 2006, ApJ, 641, 556Marois, C., Macintosh, B., Barman, T., Zuckerman, B., Song, I., Patience, J.,Lafrenière, D., & Doyon, R 2008, Science, 322, 1348Mason, B.D., Wyco ff , G.L., Hartkopf, W.I., Douglass, G.G., & Worley, C.E.2001, AJ 122, 3466Meingast, S., Alves, J., & Fürnkranz, V. 2019, A&A, 622, L13Meshkat, T., Mawet, D., Bryan, M. L., et al. 2017, AJ, 154, 245Milli, J., Vigan, A., Mouillet, D., et al. 2017, A&A, 599, A108Müller, S., Helled, R., & Mayer, L. 2018, ApJ, 854, 112Mordasini, C., Alibert, Y., & Benz, W. 2009, A&A, 501, 1139Nielsen, E. L., Liu, M. C., Wahhaj, Z., et al. 2013, ApJ, 776, 4Nielsen, E. L., De Rosa, R. J., Macintosh, B., et al. 2019, AJ, 158, 13Pecaut, M. J., & Mamajek, E. E. 2016, MNRAS, 461, 794Pourbaix, D., Tokovinin, A. A., Batten, A. H., et al. 2004, A&A, 424, 727Pueyo, L. 2016, ApJ, 824, 117Quiroga, C., Torres, A. F., & Cidale, L. S. 2010, A&A, 521, A75Rameau, J., Chauvin, G., Lagrange, A.-M., et al. 2013, A&A, 553, A60Rameau, J., Chauvin, G., Lagrange, A.-M. et al. 2013, ApJ, 772, L15Re ff ert, S., Bergmann, C., Quirrenbach, A., et al. 2015, A&A, 574, A116Rizzuto, A. C., Ireland, M. J., & Robertson, J. G. 2011, MNRAS, 416, 3108Rizzuto, A. C., Ireland, M. J., & Zucker, D. B. 2012, MNRAS, 421, L97Rizzuto, A. C., Ireland, M. J., Robertson, J. G., et al. 2013, MNRAS, 436, 1694Röser, S., Schilbach, E., Goldman, B., et al. 2018, A&A, 614, A81Rowther, S. & Meru, F. 2020, MNRAS, 496, 1598Samland, M., Bouwman, J., Hogg, D. et al. 2020, A&A, acceptedSchmid, H. M., Bazzon, A., Roelfsema, R., et al. 2018, A&A, 619, A9Schöller, M., Correia, S., Hubrig, S., et al. 2010, A&A, 522, A85Shatsky, N., & Tokovinin, A. 2002, A&A, 382, 92Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163Somers, G. & Pinsonneault, M. H. 2015, ApJ, 807, 174Somers, G., Cao, L., & Pinsonneault, M. H. 2020, ApJ, 891, 29Soummer, R., Pueyo, L., & Larkin, J. 2012, ApJL, 755, L28Stolker, T., Bonse, M. J., Quanz, S. P., et al. 2019, A&A, 621, A59Torres, C. A. O., Quast, G. R., Melo, C. H. F., et al. 2008, Handbook of StarForming Regions, Volume II, 757van Leeuwen, F. 2007, A&A, 474, 653Vigan, A., Patience, J., Marois, C., et al. 2012, A&A, 544, A9Vigan, A., Fontanive, C., Meyer, M. et al. 2020, A&A, acceptedWang, S. & Chen, X. 2019, ApJ, 877, 116Zuckerman, B. & Song, I. 2004, ARA&A, 42, 685 Article number, page 14 of 22. Janson et al.: BEAST survey
Table 1.
Observing log for BEAST first epoch data up through 2020.
Target RA Dec Date Grade a Rot. b (hh mm ss) (dd mm ss) (A / B) (deg)HIP 50847 10 22 58.15 -66 54 05.4 27 Jan 2019 B 17.6HIP 52742 10 46 57.47 -56 45 25.9 25 Jan 2019 B 22.1HIP 54767 11 12 45.21 -64 10 11.2 11 Mar 2019 A 18.5HIP 58452 11 59 10.68 -45 49 56.0 8 Apr 2018 A 31.9HIP 58901 12 04 45.25 -59 15 11.8 19 Feb 2019 B 20.3HIP 59173 12 08 05.22 -50 39 40.6 11 Mar 2019 A 26.1HIP 59747 12 15 08.72 -58 44 56.1 9 May 2019 B 22.0HIP 60009 12 18 26.25 -64 00 11.1 22 Feb 2019 A 19.0HIP 60379 12 22 49.43 -57 40 34.1 23 Feb 2019 B 22.0HIP 60710 12 26 31.76 -51 27 02.3 24 May 2019 A 26.2HIP 60823 12 28 02.38 -50 13 50.3 24 Feb 2019 B 27.3HIP 60855 12 28 22.46 -39 02 28.2 3 Jun 2019 A 45.6HIP 61257 12 33 12.19 -52 04 58.2 7 Jun 2019 A 25.7HIP 61585 12 37 11.02 -69 08 08.0 14 Mar 2019 A 17.9HIP 62058 12 43 09.18 -56 10 34.4 15 Mar 2019 A 22.2HIP 62327 12 46 22.71 -56 29 19.7 28 Mar 2019 A 22.7HIP 62434 12 47 43.27 -59 41 19.6 31 Mar 2019 A 22.5HIP 62786 12 51 56.93 -39 40 49.6 6 Feb 2020 A 42.2HIP 63003 12 54 35.62 -57 10 40.5 8 Apr 2019 A 22.8HIP 63005 12 54 36.89 -57 10 07.2 4 Mar 2019 A 21.3HIP 63945 13 06 16.70 -48 27 47.8 22 Apr 2018 A 29.3HIP 65021 13 19 43.42 -67 21 51.6 16 May 2018 A 17.3HIP 65112 13 20 37.83 -52 44 52.2 20 Mar 2019 A 24.8HIP 66454 13 37 23.47 -46 25 40.4 15 May 2018 A 30.7HIP 67464 13 49 30.28 -41 41 15.8 22 Feb 2019 A 41.0HIP 67669 13 51 49.60 -32 59 38.7 21 Feb 2019 A 72.5HIP 67703 13 52 04.85 -52 48 41.5 23 Mar 2019 A 24.9HIP 68245 13 58 16.27 -42 06 02.7 27 May 2018 A 39.1HIP 68282 13 58 40.75 -44 48 12.9 26 Mar 2019 A 34.9HIP 68862 14 06 02.77 -41 10 46.7 7 Jul 2019 A 41.0HIP 69011 14 07 40.81 -48 42 14.5 30 May 2019 B 28.8HIP 69618 14 14 57.14 -57 05 10.1 10 Jul 2019 B 21.7HIP 70300 14 23 02.24 -39 30 42.5 8 Apr 2019 A 44.9HIP 70626 14 26 49.87 -39 52 26.4 22 Apri 2018 B 43.4HIP 71352 14 35 30.42 -42 09 28.2 11 Mar 2019 A 37.5HIP 71536 14 37 53.23 -49 25 33.0 14 Mar 2019 B 27.3HIP 71865 14 41 57.59 -37 47 36.6 20 Mar 2019 A 51.3HIP 74100 15 08 39.20 -42 52 04.5 23 Mar 2019 A 36.6HIP 74657 15 15 19.64 -44 08 58.2 4 Jul 2019 A 34.1HIP 74950 15 18 56.38 -40 47 17.6 28 Apr 2019 A 37.1HIP 75304 15 23 09.35 -36 51 30.6 3 May 2019 A 51.5HIP 76048 15 31 50.23 -32 52 52.0 3 Aug 2019 A 66.3HIP 76126 15 32 55.22 -16 51 10.3 29 Jun 2019 B 74.9HIP 76591 15 38 32.64 -39 09 38.5 22 Apr 2018 B 45.2HIP 76600 15 38 39.37 -29 46 39.9 27 Jul 2019 B 90.9HIP 76633 15 39 00.06 -19 43 57.2 15 May 2018 B 100.2HIP 77562 15 50 07.08 -53 12 35.2 20 Mar 2019 B 24.5HIP 77968 15 55 22.88 -44 31 33.6 22 Jun 2018 A 34.1HIP 78104 15 56 53.08 -29 12 50.7 29 Jul 2019 A 107.4HIP 78207 15 58 11.37 -14 16 45.7 8 Apr 2018 A 46.3HIP 78655 16 03 24.19 -38 36 09.2 8 Apr 2019 B 41.2HIP 78918 16 06 35.55 -36 48 08.3 23 May 2019 A 49.4HIP 78968 16 07 14.93 -17 56 09.7 31 Jul 2019 B 77.1HIP 79044 16 08 04.38 -36 13 54.6 21 Mar 2019 A 53.9HIP 79404 16 12 18.20 -27 55 34.9 9 May 2019 A 124.7HIP 80142 16 21 27.03 -48 11 19.0 26 Jun 2019 A 22.8HIP 80208 16 22 28.00 -49 34 20.5 9 Jul 2019 B 27.7HIP 80569 16 27 01.43 -18 27 22.5 5 Jun 2019 A 83.5HIP 80911 16 31 22.93 -34 42 15.7 13 Aug 2018 B 63.0HIP 81208 16 35 13.84 -35 43 28.7 5 Aug 2019 A 49.6HIP 81316 16 36 28.67 -40 18 10.9 29 Jun 2019 B 42.3HIP 81472 16 38 26.29 -43 23 54.3 10 Jul 2019 B 34.7HIP 81891 16 43 38.72 -32 06 21.4 3 May 2019 A 57.7HIP 81914 16 43 54.08 -41 06 48.0 23 May 2019 A 40.4HIP 81972 16 44 42.59 -40 50 22.8 7 Aug 2019 B 39.6HIP 82514 16 51 52.23 -38 02 50.6 14 May 2018 A 48.8HIP 82545 16 52 20.15 -38 01 03.1 23 Apr 2018 A 51.1 a ESO rating of ambient conditions, see text for details. b Field rotation between first and last useful exposure (pupil stabilized mode).
Article number, page 15 of 22 & Aproofs: manuscript no. main
Table 2.
Subgroup membership of the BEAST sample.
Target Alt name RA Dec Plx σ plx µ RA µ Dec
Region Prob.(hh mm ss) (dd mm ss) (mas) (mas) (mas / yr) (mas / yr) (%)HIP 50847 L Car 10 22 58.15 -66 54 05.4 8.16 0.20 -22.23 12.17 LCC 79.9HIP 52742 HR 4221 10 46 57.47 -56 45 25.9 4.98 0.23 -20.71 0.57 LCC 0.0HIP 54767 HD 97583 11 12 45.21 -64 10 11.2 10.59 0.14 -47.32 5.46 LCC 3.9HIP 58452 HD 104080 11 59 10.68 -45 49 56.0 7.52 0.07 -30.43 -9.03 LCC 46.8HIP 58901 HD 104900 12 04 45.25 -59 15 11.8 9.14 0.08 -38.22 -12.04 LCC 99.3HIP 59173 HR 4618 12 08 05.22 -50 39 40.6 9.85 0.32 -34.35 -11.51 LCC 99.1HIP 59747 δ Cru 12 15 08.72 -58 44 56.1 11.62 0.96 -37.29 -9.97 LCC 99.6HIP 60009 ζ Cru 12 18 26.25 -64 00 11.1 10.49 0.33 -34.96 -6.24 LCC 99.9HIP 60379 HR 4706 12 22 49.43 -57 40 34.1 9.69 0.16 -37.25 -12.11 LCC 95.1HIP 60710 G Cen 12 26 31.76 -51 27 02.3 7.85 0.37 -31.42 -10.41 LCC 99.0HIP 60823 σ Cen 12 28 02.38 -50 13 50.3 8.88 0.33 -32.11 -13.81 LCC 98.3HIP 60855 u Cen 12 28 22.46 -39 02 28.2 7.40 0.24 -26.89 -7.32 LCC 4.1HIP 61257 HD 109195 12 33 12.19 -52 04 58.2 7.99 0.07 -31.40 -12.57 LCC 99.1HIP 61585 α Mus 12 37 11.02 -69 08 08.0 8.62 0.83 -29.37 -11.38 LCC 89.1HIP 62058 HR 4834 12 43 09.18 -56 10 34.4 7.90 0.06 -31.56 -12.30 LCC 99.7HIP 62327 HD 110956 12 46 22.71 -56 29 19.7 10.03 0.34 -33.53 -14.97 LCC 99.4HIP 62434 β Cru 12 47 43.27 -59 41 19.6 11.71 0.98 -42.97 -16.18 LCC 99.1HIP 62786 HR 4879 12 51 56.93 -39 40 49.6 6.89 0.08 -25.73 -18.86 UCL 78.5HIP 63003 µ .01 Cru 12 54 35.62 -57 10 40.5 9.63 0.36 -31.09 -14.51 LCC 98.6HIP 63005 µ .02 Cru 12 54 36.89 -57 10 07.2 8.95 0.23 -28.16 -10.34 LCC 97.3HIP 63210 H Cen 12 57 04.35 -51 11 55.5 7.93 0.19 -30.57 -15.50 LCC 98.9HIP 63945 f Cen 13 06 16.70 -48 27 47.8 7.58 0.33 -30.05 -14.55 LCC 95.1HIP 64004 ξ
02 Cen 13 06 54.64 -49 54 22.5 6.70 0.29 -26.36 -10.58 LCC 95.9HIP 65021 HD 115583 13 19 43.42 -67 21 51.6 5.76 0.05 -23.43 -13.86 LCC 0.0HIP 65112 V964 Cen 13 20 37.83 -52 44 52.2 8.17 0.17 -30.44 -16.30 LCC 97.0HIP 66454 HR 5121 13 37 23.47 -46 25 40.4 7.07 0.23 -29.53 -18.05 UCL 73.5HIP 67464 ν Cen 13 49 30.28 -41 41 15.8 10.05 0.47 -27.86 -18.08 UCL 95.9HIP 67669 VV983 Cen 13 51 49.60 -32 59 38.7 11.10 0.43 -34.70 -27.91 UCL 96.7HIP 67703 HD 120642 13 52 04.85 -52 48 41.5 10.48 0.16 -39.04 -27.43 LCC 82.5HIP 68245 φ Cen 13 58 16.27 -42 06 02.7 8.51 0.36 -25.55 -17.27 UCL 96.3HIP 68282 υ
01 Cen 13 58 40.75 -44 48 12.9 8.73 0.59 -24.02 -21.84 UCL 96.5HIP 68862 χ Cen 14 06 02.77 -41 10 46.7 7.75 0.38 -24.60 -20.19 UCL 99.1HIP 69011 HD 123247 14 07 40.81 -48 42 14.5 10.13 0.07 -34.20 -28.27 UCL 76.8HIP 69618 HD 124367 14 14 57.14 -57 05 10.1 7.89 0.23 -25.10 -20.19 UCL 87.3HIP 70300 a Cen 14 23 02.24 -39 30 42.5 10.13 0.44 -24.95 -18.61 UCL 96.5HIP 70626 HR 5400 14 26 49.87 -39 52 26.4 6.95 0.10 -20.63 -20.86 UCL 99.8HIP 71352 η Cen 14 35 30.42 -42 09 28.2 10.67 0.21 -34.73 -32.72 UCL 99.6HIP 71353 HR 5439 14 35 31.48 -41 31 02.8 9.38 0.34 -22.17 -20.93 UCL 99.4HIP 71453 HD 128207 14 36 44.13 -40 12 41.7 7.00 0.16 -23.71 -22.32 UCL 99.8HIP 71536 ρ Lup 14 37 53.23 -49 25 33.0 10.19 0.54 -28.57 -26.48 UCL 99.2HIP 71860 α Lup 14 41 55.76 -47 23 17.5 7.02 0.17 -20.94 -23.67 UCL 99.9HIP 71865 b Cen 14 41 57.59 -37 47 36.6 10.24 0.64 -31.52 -31.50 UCL 99.8HIP 73266 HD 132094 14 58 24.26 -37 21 44.9 5.66 0.13 -19.46 -22.69 UCL 94.5HIP 73624 HD 132955 15 02 59.28 -32 38 35.9 6.58 0.18 -19.74 -21.57 UCL 98.9HIP 74100 HR 5625 15 08 39.20 -42 52 04.5 6.76 0.09 -20.87 -21.65 UCL 99.9HIP 74449 e Lup 15 12 49.59 -44 30 01.5 7.64 0.97 -22.02 -22.16 UCL 99.4HIP 74657 HD 135174 15 15 19.64 -44 08 58.2 6.43 0.12 -19.37 -21.75 UCL 99.8HIP 74752 HD 135454 15 16 37.15 -42 22 12.6 7.45 0.07 -22.16 -27.38 UCL 99.9HIP 74950 VGG Lup 15 18 56.38 -40 47 17.6 6.97 0.18 -20.07 -21.58 UCL 99.9HIP 75141 δ Lup 15 21 22.33 -40 38 51.0 8.72 0.47 -17.28 -23.15 UCL 99.6HIP 75304 φ
02 Lup 15 23 09.35 -36 51 30.6 6.28 0.20 -18.24 -20.72 UCL 99.6HIP 75647 HR 5736 15 27 18.13 -36 46 03.2 7.11 0.22 -19.55 -23.58 UCL 99.9HIP 76048 HR 5753 15 31 50.23 -32 52 52.0 6.31 0.12 -17.75 -22.15 UCL 99.1HIP 76126 ζ
04 Lib 15 32 55.22 -16 51 10.3 4.53 0.19 -11.50 -10.85 US 0.0HIP 76591 HR 5805 15 38 32.64 -39 09 38.5 6.85 0.05 -19.81 -23.90 UCL 99.9HIP 76600 τ Lib 15 38 39.37 -29 46 39.9 4.14 0.76 -11.08 -22.88 UCL 2.9HIP 76633 HD 139486 15 39 00.06 -19 43 57.2 6.19 0.06 -15.30 -18.22 US 76.7HIP 77562 HD 141168 15 50 07.08 -53 12 35.2 10.28 0.10 -27.59 -39.27 UCL 95.2HIP 77968 HD 142256 15 55 22.88 -44 31 33.6 5.50 0.06 -15.66 -19.65 UCL 80.7
Article number, page 16 of 22. Janson et al.: BEAST survey
Table 2. continued.
Target Alt name RA Dec Plx σ plx µ RA µ Dec
Region Prob.(hh mm ss) (dd mm ss) (mas) (mas) (mas / yr) (mas / yr) (%)HIP 78104 ρ Sco 15 56 53.08 -29 12 50.7 7.49 0.41 -18.08 -24.39 US 78.5HIP 78168 HD 142883 15 57 40.46 -20 58 59.1 6.39 0.10 -10.13 -21.75 US 99.4HIP 78207 48 Lib 15 58 11.37 -14 16 45.7 7.52 0.25 -14.92 -16.42 US 81.9HIP 78324 HD 143022 15 59 30.88 -40 51 54.6 6.02 0.05 -16.51 -20.31 UCL 97.4HIP 78384 η Lup 16 00 07.33 -38 23 48.2 5.99 0.41 -18.64 -28.46 UCL 97.3HIP 78655 HR 5967 16 03 24.19 -38 36 09.2 9.23 0.47 -18.49 -28.15 UCL 99.8HIP 78702 HD 143956 16 04 00.24 -19 46 02.9 6.56 0.07 -9.89 -21.48 US 99.8HIP 78918 θ Lup 16 06 35.55 -36 48 08.3 8.17 0.38 -20.10 -33.42 UCL 94.4HIP 78933 ω Sco 16 06 48.43 -20 40 09.1 7.06 0.40 -7.91 -21.06 US 99.9HIP 78968 HD 144586 16 07 14.93 -17 56 09.7 6.62 0.08 -8.77 -21.27 US 99.7HIP 79044 HD 144591 16 08 04.38 -36 13 54.6 7.29 0.05 -16.88 -28.04 UCL 99.7HIP 79404 c02 Sco 16 12 18.20 -27 55 34.9 6.65 0.36 -11.81 -23.76 US 97.8HIP 80142 HD 147001 16 21 27.03 -48 11 19.0 5.56 0.06 -12.75 -21.97 UCL 11.7HIP 80208 HD 147152 16 22 28.00 -49 34 20.5 5.30 0.44 -11.32 -25.75 UCL 3.2HIP 80569 χ Oph 16 27 01.43 -18 27 22.5 8.18 0.31 -4.97 -21.75 US 93.7HIP 80911 N Sco 16 31 22.93 -34 42 15.7 7.45 0.56 -11.01 -18.08 UCL 96.8HIP 81208 HD 149274 16 35 13.84 -35 43 28.7 6.75 0.07 -9.61 -25.70 UCL 94.9HIP 81266 τ Sco 16 35 52.95 -28 12 57.7 6.88 0.53 -9.89 -22.83 US 77.5HIP 81316 HD 149425 16 36 28.67 -40 18 10.9 5.23 0.06 -11.98 -20.79 UCL 36.8HIP 81472 VV1003 Sco 16 38 26.29 -43 23 54.3 5.12 0.11 -11.05 -20.25 UCL 28.1HIP 81474 HD 149914 16 38 28.65 -18 13 13.7 6.30 0.08 -13.91 -20.38 UCL 7.5HIP 81891 HR 6211 16 43 38.72 -32 06 21.4 6.41 0.07 -8.79 -23.17 UCL 93.9HIP 81914 HD 150591 16 43 54.08 -41 06 48.0 5.66 0.08 -12.95 -21.76 UCL 58.6HIP 81972 HD 150742 16 44 42.59 -40 50 22.8 5.84 0.11 -12.44 -21.67 UCL 85.8HIP 82514 µ .01 Sco 16 51 52.23 -38 02 50.6 a µ .02 Sco 16 52 20.15 -38 01 03.1 7.92 0.55 -9.98 -19.88 UCL 96.3 a Unreliable value for HIP 82514; see Sect. 2.2
Table 3.
Age estimations for the sample.
Target CMS age a MAP age b SG age c (Myr) (Myr) (Myr)HIP 50847 27.3 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Article number, page 17 of 22 & Aproofs: manuscript no. main
Table 3. continued.
Target CMS age a MAP age b SG age c (Myr) (Myr) (Myr)HIP 67669 — — 16 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Article number, page 18 of 22. Janson et al.: BEAST survey
Table 3. continued.
Target CMS age a MAP age b SG age c (Myr) (Myr) (Myr) a Age based on co-moving stars; see text. b Age based on the map in Pecaut & Mamajek (2016). c Age of the closest matching Sco-Cen subgroup.
Table 4.
Summary of known companions and co-moving stars
Target SB / EB a Reference AO / INT b Reference CMS c HIP 50847 SB2 Quiroga et al. (2010) AO This work 18HIP 52742 — — — — 1HIP 54767 — — — — 4HIP 58452 — — — — 21HIP 58901 — — — — 172HIP 59173 SB2 Chini et al. (2012) AO This work 52HIP 59747 SB2 Chini et al. (2012) — — 0HIP 60009 SB2 Chini et al. (2012) AO This work 14HIP 60379 — — — — 140HIP 60710 — — — — 12HIP 60823 — — INT Rizzuto et al. (2013) 44HIP 60855 — — — — 0HIP 61257 — — — — 202HIP 61585 SB2 Chini et al. (2012) INT Rizzuto et al. (2013) 0HIP 62058 — — — — 214HIP 62327 SB1 Chini et al. (2012) — — 57HIP 62434 SB1 Pourbaix et al. (2004) AO This work 11HIP 62786 — — — — 0HIP 63003 — — — — 28HIP 63005 — — AO This work 93HIP 63210 SB2 Chini et al. (2012) — — 71HIP 63945 SB1 Chini et al. (2012) INT + AO See note d / EB Pourbaix et al. (2004) — — 0HIP 67669 SB1 Pourbaix et al. (2004) INT Schöller et al. (2010) 0HIP 67703 — — — — 12HIP 68245 — — — — 0HIP 68282 — — — — 0HIP 68862 — — — — 0HIP 69011 — — — — 17HIP 69618 — — — — 10HIP 70300 — — — — 0HIP 70626 — — — — 111HIP 71352 SB2 Chini et al. (2012) — — 4HIP 71353 — — — — 0HIP 71453 — — — — 6HIP 71536 — — INT Rizzuto et al. (2013) 0HIP 71860 — — — — 4HIP 71865 — — INT Rizzuto et al. (2013) 0HIP 73266 — — — — 0HIP 73624 — — AO This work 7HIP 74100 — — AO Shatsky & Tokovinin (2002) 100HIP 74449 SB1 Pourbaix et al. (2004) — — 0HIP 74657 — — — — 54HIP 74752 — — — — 46HIP 74950 SB2 / EB Budding et al. (2015) — — 58HIP 75141 — — — — 0HIP 75304 — — INT Rizzuto et al. (2013) 15HIP 75647 SB1 Pourbaix et al. (2004) — — 31HIP 76048 — — — — 21
Article number, page 19 of 22 & Aproofs: manuscript no. main
Table 4. continued.
Target SB / EB a Reference AO / INT b Reference CMS c HIP 76126 — — — — 0HIP 76591 — — — — 136HIP 76600 SB2 Pourbaix et al. (2004) INT Rizzuto et al. (2013) 0HIP 76633 — — — — 2HIP 77562 — — — — 8HIP 77968 — — — — 42HIP 78104 SB1 Pourbaix et al. (2004) — — 0HIP 78168 SB2 / EB David et al. (2019) — — 54HIP 78207 — — — — 0HIP 78324 — — — — 27HIP 78384 — — — — 0HIP 78655 — — — — 0HIP 78702 — — — — 97HIP 78918 — — — — 0HIP 78933 — — — — 0HIP 78968 — — — — 91HIP 79044 — — — — 228HIP 79404 SB1 Pourbaix et al. (2004) — — 0HIP 80142 — — — — 10HIP 80208 — — — — 0HIP 80569 SB1 Pourbaix et al. (2004) — — 0HIP 80911 — — — — 0HIP 81208 — — — — 61HIP 81266 — — INT Rizzuto et al. (2013) 0HIP 81316 — — — — 36HIP 81472 — — — — 13HIP 81474 — — — — 0HIP 81891 — — — — 67HIP 81914 — — — — 192HIP 81972 — — — — 236HIP 82514 SB2 / EB Budding et al. (2015) — — 0HIP 82545 — — — — 0 a Spectroscopic (SB) and eclipsing (EB) binary companions. b Companions identified with AO or interferometry (INT). c Number of identified candidate co-moving stars. Most are not physically bound. d Interferometric companion from Rizzuto et al. (2013) and AO companion from Shatsky & Tokovinin (2002).
Table 5.
Fundamental properties of the BEAST sample.
Target Dist SpT Mass App. V App. K Disc? Ref. a (pc) M sun (mag) (mag)HIP 50847 122.6 B8V 2.9 4.99 5.314HIP 52742 200.8 B8.5III 2.7 5.23 5.225 Be-type A17HIP 54767 94.5 B8V 2.9 5.22 5.418HIP 58452 133.1 B8 / B9V 2.7 6.34 6.529HIP 58901 109.4 B9V 2.5 6.20 6.278HIP 59173 101.5 B6III 3.7 4.47 4.871 Excess R12HIP 59747 86.0 B2IV 7.8 2.78 3.532HIP 60009 95.3 B2.5V 6.8 4.05 4.535HIP 60379 103.2 B9V 2.5 5.37 5.636HIP 60710 127.3 B3V 5.9 4.81 5.234 Excess C16HIP 60823 112.7 B2V 7.8 3.91 4.478HIP 60855 135.1 B8V 2.9 5.45 5.541 Excess R12HIP 61257 125.2 B9V 2.5 6.55 6.612HIP 61585 116.0 B2IV-V 7.8 2.68 3.250HIP 62058 126.6 B7.5V 3.1 5.99 6.173HIP 62327 99.7 B3V 5.9 4.61 5.056HIP 62434 85.4 B0.5IV 13 1.25 1.978HIP 62786 145.1 B8V 2.9 5.97 6.223HIP 63003 103.9 B2IV-V 7.8 4.04 4.529
Article number, page 20 of 22. Janson et al.: BEAST survey
Table 5. continued.
Target Dist SpT Mass App. V App. K Disc? Ref. a (pc) M sun (mag) (mag)HIP 63005 111.7 B5V 4.2 5.20 5.313 Excess R12HIP 63210 126.1 B2IV 7.8 5.16 5.342HIP 63945 131.9 B5V 4.2 4.69 5.045HIP 64004 149.2 B1.5V 9 4.27 4.806HIP 65021 173.5 B9V 2.5 7.25 7.223HIP 65112 122.5 B6V 3.7 5.45 5.802HIP 66454 141.4 B8V 2.9 5.89 6.136HIP 67464 99.5 B2IV 7.8 3.39 4.240HIP 67669 90.1 B5III 4.2 4.53 4.971 Excess R12HIP 67703 95.4 B9V 2.5 5.25 5.515HIP 68245 117.5 B2IV 7.8 3.81 4.491 Excess R12HIP 68282 114.6 B2IV-V 7.8 3.87 4.469HIP 68862 129.0 B2V 7.8 4.34 4.931HIP 69011 98.8 B9V 2.5 6.43 6.434 Excess C16HIP 69618 126.7 B4V 5.1 5.07 4.766 Exc + Be R12 + A17HIP 70300 98.7 B7III 3.3 4.42 4.923 Excess R12HIP 70626 143.9 B7IV 3.3 6.35 6.592HIP 71352 93.7 B2V 7.8 2.31 2.750 Exc + Be R12 + A17HIP 71353 106.6 B7V 3.3 5.87 6.030HIP 71453 142.8 B8V 2.9 5.74 6.015 Excess C16HIP 71536 98.2 B5V 4.2 4.05 4.507HIP 71860 142.5 B1.5III 9 2.28 2.668HIP 71865 97.6 B3V 5.9 4.00 4.487HIP 73266 176.7 B9V 2.5 7.25 7.281HIP 73624 151.9 B3V 5.9 5.43 5.734HIP 74100 147.9 B7V 3.3 5.82 6.097HIP 74449 130.9 B3IV / V 5.9 4.81 5.283HIP 74657 155.6 B9IV 2.5 6.73 6.945HIP 74752 134.1 B9.5 2.4 6.75 6.834 Excess C16HIP 74950 143.4 B7V 3.3 5.59 6.212HIP 75141 114.7 B1.5IV 9 3.19 3.960HIP 75304 159.2 B4V 5.1 4.54 4.943 Excess R12HIP 75647 140.7 B4V 5.1 5.45 5.840HIP 76048 158.5 B7V 3.3 6.49 6.247HIP 76126 220.7 B3V 5.9 5.49 5.891HIP 76591 146.0 B9V 2.5 6.58 6.763HIP 76600 241.4 B2.5V 6.8 3.64 4.120HIP 76633 161.5 B9V 2.5 7.63 7.492 Excess L12HIP 77562 97.2 B9V 2.5 5.78 5.945HIP 77968 182.0 B8V 2.9 6.96 6.987HIP 78104 133.5 B2IV-V 7.8 3.86 4.457HIP 78168 156.4 B3V 5.9 5.84 5.734HIP 78207 133.0 B8Ia / Iab 2.9 4.87 4.591 Exc + Be C09HIP 78324 166.1 B9V 2.5 8.18 7.488HIP 78384 167.1 B2.5IV 6.8 3.41 4.090HIP 78655 108.4 B6IV 3.7 4.89 5.271HIP 78702 152.5 B9V 2.5 7.77 7.237 Excess C16HIP 78918 122.3 B2.5V 6.8 4.21 4.699HIP 78933 141.7 B1V 10 3.97 4.009HIP 78968 151.0 B9V 2.5 7.81 7.401 Excess C16HIP 79044 137.1 B9V 2.5 6.74 6.913HIP 79404 150.3 B2V 7.8 4.57 4.976HIP 80142 179.8 B7V 3.3 6.50 6.619HIP 80208 188.8 B3V 5.9 5.32 5.397HIP 80569 122.3 B2V 7.8 4.43 2.885 Exc + Be L12 + A17HIP 80911 134.2 B2III-IV 7.8 4.23 4.693HIP 81208 148.1 B9V 2.5 6.64 6.768HIP 81266 145.3 B0V 16 2.81 3.705HIP 81316 191.2 B9V 2.5 7.10 6.714 Excess R12HIP 81472 195.1 B2V 7.8 5.83 5.933
Article number, page 21 of 22 & Aproofs: manuscript no. main
Table 5. continued.
Target Dist SpT Mass App. V App. K Disc? Ref. a (pc) M sun (mag) (mag)HIP 81474 158.8 B9.5IV 2.4 6.75 5.690 Excess C16HIP 81891 156.1 B8V 2.9 6.46 6.630 Excess L14HIP 81914 176.7 B6.5V 3.5 6.13 6.293HIP 81972 171.2 B3V 5.9 5.64 5.843 Exc + Be C16 + A17HIP 82514 b a References. A17: Arcos et al. (2017). C09: Carpenter et al. (2009). C16: Cotten & Song (2016).L12: Luhman & Mamajek (2012). L14: Liu et al. (2014). R12: Rizzuto et al. (2012). b Unreliable value for HIP 82514; see Sect. 2.2
Table 6.
Properties of binary companions in the BEAST data.
Target Pair Date ρ (mas) σ ρ (mas) θ (deg) σ θ (deg) Disc. a HIP 50847 AB 27 Jan 2019 2215.1 3.0 352.45 0.30 TWHIP 59173 AB 11 Mar 2019 1275.2 2.8 129.88 0.09 TWHIP 60009 AB 22 Feb 2019 148.2 0.7 170.89 0.11 TWHIP 62434 AB 31 Mar 2019 4176.1 3.1 121.84 0.05 TWHIP 63005 AB 4 Mar 2019 274.0 1.1 172.36 0.12 TWHIP 63945 AC b