Mem.S.A.It.Vol.75,282 (cid:13)c SAIt 2008 Memoriedella Globular clusters seen by Gaia E. Pancino1,2,M. Bellazzini1,andS. Marinoni2,3 1 IstitutoNazionalediAstrofisica–OsservatorioAstronomicodiBologna,ViaRanzani1, I-40127Bologna,Italy 3 2 ASIScienceDataCenter,I-00044Frascati,Italy 1 3 IstitutoNazionalediAstrofisica–OsservatorioAstronomicodiRoma,ViadiFrascati33, 0 2 Monteporzio(RM),Italy e-mail:[email protected] n a J Abstract.Wepresentasimulationoftwelveglobularclusterswithdifferentconcentration, 8 distance,andbackgroundpopulation,whosepropertiesaretransformedintoGaiaobserv- 1 ableswiththehelpofthelatesGaiascienceperformancesprescriptions.Weadoptsimplified crowdingreceipts,basedonfiveyearsofsimulationsperformedbyDPAC(DataProcessing ] A and Analysis Consortium) scientists, to explore the effect of crowding and to give a ba- sic idea of what will be made possible by Gaia in the filed of Galactic globular clusters G observations. . h Keywords.Spacevehicles:instruments—Galaxy:globularclusters—Methods:miscel- p laneous - o r t s 1. Introduction external parts. With all Gaia scientistists busy a inpre-launchactivities,weattemptheretouse [ GaiaisacornerstoneESAastrometricmission the results of 5–10 years of stellar blending whichisgoingtobelaunchedinOctober2013. 1 simulations with GIBIS (Babusiaux 2005) to v Itwillprovide6Dpositionandvelocityinfor- derive simplified receipts, and to give a basic 6 mationonGalacticstars,solarsystemobjects, ideaonwhatwillbemadepossiblebyGaiain 6 unresolvedgalaxies,withexquisitequalityand thefieldofGalacticglobularclustersobserva- 3 additional astrophysical information such as tions. 4 E(B–V),astrophysicalparameters,objectclas- 1. sification, chemical tagging, for objects down 0 to V(cid:39)20 mag. A deep discussion about the 2. Simulations 3 Gaia expected scientific harvest can be found 1 inmanypapers(seee.g.,Mignard2005). TwoclustersweresimulatedwiththeMcluster v: Gaia will certainly be able to provide in- code (Ku¨pper et al. 2011) — including the i terestingdataforglobularclustersaswell,but prescriptions by Hurley et al. (2000) — dif- X crowdingisthegreatunknown,itsknowledge fering only in their concentration parameter, r deciding whether we will be able to pierce one having c=1.0 and r =5 pc and the other a h throughthecentralregionsorbelimitedtothe c=2.5 and r =4 pc. Both clusters were spher- h ical, non-rotating, and did not include bina- Sendoffprintrequeststo:E.Pancino ries.Theunderlyingstellarpopulationwasold Pancino:GGCseenbyGaia 283 Fig.1.AGIBIS(Babusiaux2005)simulationcomputedbyG.Giuffrida(2012,privatecommunication)of aglobularclusterobservedthroughtheGaiaspectro-photometers.Eachstarhasacygnus-likeshape,and greenrectanglesarethewindowsassignedtobrightstars(G≤20mag)bytheon-boarddetectionalgorithm. The750000windowslimitimposesthatnonewwindowisassignedtoanewstar(enteringfromtheleft) untiloneofthepresentlyassignedwindowsisfreedbyastartransitingoutofthefield(fromtherightside). Morecrowdedregionshaveahigherprobabilityofhavingawindowassigned,thus,theclustercenterhas morewindowsassignedthanthetwoexternalregionsaboveandbelowit.Thisisratherunusualcompared totypicalground-basedobservationsofcrowdedregions,andhasimplicationsforthecompletenesslevels ofthephotometry(courtesyofG.Giuffrida). (12Gyr),metal-poor(Z=0.003),andtheother CCDsreadinTDI(TimeDelayedIntegration) clusterpropertieswererathertypicaloftheav- mode, as described in many Gaia technical erage Milky Way globular clusters (for exam- documents (but see also Cacciari 2011; Jordi ple:M =–7.6mag,M =3.6105 M ,Kroupa 2011; Pancino 2012). Crowding thus has an V tot (cid:12) IMF). Each cluster was projected at three dis- impact which deserves in depth studies to tances (5, 10, 15 kpc) and onto two back- be fully understood. Many groups used the grounds, simulated with the Besanc¸on model various Gaia simulators to better understand inahalodirection(l=150,b=80)andacrowded crowding both for deblending and for image bulge field (l=5,b=5). Thus, in total, we sim- reconstrunction.Hereweusetheirworktode- ulated twelve clusters with different combi- rivesimplifiedreceiptsthatcanbeusedinour nations of concentration, distance, and back- simulations of globular clusters. In particular, groundcrowding. weclassifyblendsasfollows: We transformed the simulated properties of each star into Gaia observables and their uncertainties (injecting Gaussian noise) using Hard blends (or classical blends). They the latest version of the Gaia science perfor- occur when two or more stars are closer than mances1andrelatedpublishedtransformations the FWHM in the astrometric field (AF) of (e.g.,Jordietal.2010;Kordopatisetal.2011). Gaia ((cid:39)0.53”); they could in principle be de- blendedusinginformationowingtothediffer- entorientationofeachGaiatransit,andtothe 3. Gaiaandcrowding radial velocity variations in the radial veloc- ity spectrometer (RVS), but not having suffi- Gaiaisacomplexinstrument,withalargefo- cient technical literature, for the moment we cal plane hosting three instruments and many justflaggedthesecasesinoursimulations;we 1 http://www.rssd.esa.int willtreatthematalaterstage. 284 Pancino:GGCseenbyGaia Fig.2.Thecolor-magnitudediagraminGaiacolorsinanareaof0.7×0.7degaroundthesimulated“diffi- cult”cluster(seetext).Blacksymbolsinbothpanelsrefertofieldstars,yellowsymbolstodeblendedstars (seetextforadescription),greensymbolstodecontaminatedstars,andbluesymbolstocleanstars.The left panel shows the full catalogue, while the righ panel shows only bona-fide cluster members selected by proper motions and distance from the cluster center; in this panel two cluster stars were accidentally rejected and are colored in magenta, while only a handful of the otherwise dominating field population remain after the selection. The number of well measured member stars is small, but they all have space qualitymeasurements(oftheorderofammagprecisionandafew%calibrationforthebrightstars). Blends. Blends occur when two or more Contaminated stars. When two stars are starsarecloserthantheshortsideoftheread- not blended, but are closer than the long side out window assigned in the AF and in the oftheassignedwindow(3.54”inBP/RP),they BP/RP instruments (Blue and Red spectro- will be assigned the same window in some Photometers),whichis2.12”.Inthiscase,the transit and separarate windows in some other stars will always be assigned the same win- transits, depending on the orientation of the dow (thus be compressed to one single 1D Gaiascandirectiononthesky.Thustheywill spectrumbeforegroundtransmission),nomat- beeasiertodisentanglethanblends.However, ter the orientation on the sky. However, the brightstars(V≤15mag)willhaveenoughflux spectra will overlap differently for different outside their windows to contaminate neigh- orientations in the sky, and will have differ- bouringstarssoavariable“size”dependingon ent radial velocities, measurable on RVS (for theirmagnitudeneedstobedefined(Marrese, V<17 mag). All these effects can be modeled 2008,privatecommunication)totakethisinto andthestarswilleffectivelybedeblendedwith account. In general, various experiments on residuals of (cid:39)3–5% according to a test made two stars in isolation showed that they can be byG.GiuffridawithGIBIS(Babusiaux2005). decontaminated with residuals better than 1– Thetestwasmadewithouttakingintoaccount 3%.WeconservativelydegradedtheGaiasci- CTI (Charge transfer inefficiencies), however, ence performances of (de)contaminated stars sowedegradedtheGaiascienceperformances by 5%, proportionally to the flux contamina- of(de)blendedstarsby10%,proportionallyto tion. thefluxcontamination. Pancino:GGCseenbyGaia 285 Table1.Systemicpropertiesoftwoclusters(seetext) Property Easycluster Difficultcluster true(input)value #ofstars 16838 3513 — µ –4998.7pm0.8µas/yr –4993±3µas/yr –5000µas/yr RA µ –5000.2pm0.7µas/yr -4994±3µas/yr –5000µas/yr Dec π 199.7±0.7µas 101.2±1.4µas 200/100µas D 5.007±0.007kpc 9.997±0.017kpc 5/10kpc 4. Resultsandconclusions difficult cluster has c=2.5, d=10 kpc, and bulge-likebackgoundandisalsoshownin In general, as was largely expected, we see Figure2)2. that crowding has a large effect on the central areas of clusters, generally making Gaia per- Acknowledgements. Thisworkusessimulateddata formances rather poor (for a space telescope) providedbytheSimulationUnit(CU2)oftheGaia inside the half-light radius of a cluster. Also, DataProcessingAnalysisConsortium(DPAC),run Gaia does not go very deep (V<20 mag) and withGIBISatCNES(Centrenationald’e´tudesspa- thus clusters farther than 15 kps are not sam- tiales). pled down to their turnoff point. However, for thosestarsthataremeasured,Gaiagrants: References – an excellent membership probability as- Babusiaux, C. 2005, The Three-Dimensional sessment using its superb proper motions UniversewithGaia,576,417 and — for the bright stars — the RVS ra- Cacciari, C. 2011, EAS Publications Series, dial velocities; most samples obtained by 45,155 Gaia will suffer from very low field con- Hurley,J.R.,Pols,O.R.,&Tout,C.A.2000, taminationeveninthemostcrowdedenvi- MNRAS,315,543 ronments; see Figure 2 for a difficult case Jordi, C., Gebran, M., Carrasco, J. M., et al. diagram,cleanedwithpropermotions; 2010,A&A,523,A48 – space-quality photometry (see Figure 2), Jordi, C. 2011, EAS Publications Series, 45, with mmag precision down to approxi- 149 mately16–17magatleast,andaphotomet- Mignard,F.2005,AstrometryintheAgeofthe riccalibrationofafewpercentaccuracyat Next Generation of Large Telescopes, 338, most, for a number of stars ranging from 15 afewhundred(forthedifficultclusters)to Pancino,E.2012,arXiv:1207.1616 a few tens of thousands (for a few tens of Kordopatis,G.,Recio-Blanco,A.,deLaverny, clusters); P.,etal.2011,A&A,535,A106 – exquisite proper motions for the above Ku¨pper,A.H.W.,Maschberger,T.,Kroupa,P., stars, with µas/yr errors, at least down &Baumgardt,H.2011,MNRAS,417,2300 to 16–17 mag (or (cid:39)300 µas/yr down to 20mag),andparallaxeswithsimilarerrors, complementedby1–10km/sradialveloci- tiesforstarsdowntoV(cid:39)17mag; – statistical distances and systemic proper motionswithunprecedentedaccuracy(see 2 The distances in this paper are based only Table 1, where the easy cluster has c=1.0, on parallaxes, but of course there will be many d=5 kpc, and halo-like background; the RRLyraeinglobularclustersaswell.