The Proper Motion of the Globular Cluster NGC 6553 and of Bulge Stars with HST

16 Pages·1.9 MB·English

by  M. Zoccali

The Proper Motion of the Globular Cluster NGC 6553 and of
Bulge Stars with HST1
M. Zoccali2, A. Renzini2, S. Ortolani3, E. Bica4, B. Barbuy5
1
0
0
ABSTRACT
2
n WFPC2 images obtained with the Hubble Space telescope 4.16 years apart have allowed us
a to measure the proper motion of the metal rich globular cluster NGC 6553 with respect to the
J backgroundbulgestars. Withaspacevelocityof(Π, Θ, W)=(-3.5,230,-3)kms−1,NGC6553

6 follows the mean rotation of both disk and bulge stars at a Galactocentric distance of 2.7 kpc.
1
While the kinematics of the cluster is consistent with either a bulge or a disk membership, the

virtualidentity of its stellar populationwith that of the bulge cluster NGC 6528makesits bulge
2
v membership more likely. The astrometric accuracy is high enough for providing a measure of
0 the bulge proper motion dispersion and confirming its rotation. A selection of stars based on
0 thepropermotionsproducedanextremelywelldefinedclustercolor-magnitudediagram(CMD),
2 essencially free of bulge stars. The improved turnoff definition in the decontaminated CMD
1 confirms an old age for the cluster (∼ 13 Gyr) indicating that the bulge underwent a rapid
0
chemical enrichment while being built up at in the early Universe. An additional interesting
1
0 feature of the cluster color-magnitude diagram is a significant number of blue stragglers stars,
/ whose membership in the cluster is firmly established from their proper motions.
h
p
Subject headings: Astrometry; Globular clusters: individual: NGC 6553; The Galaxy: kinematics and
-
o dynamics;
r
t 1. INTRODUCTION 1989). These clusters are actually strongly con-
s
a centrated towards the Galactic center, with most
: The globular cluster system of the Milky Way
v ofthembeingphysicallyinsidetheGalacticbulge,
i hasaslightlybimodalmetallicitydistribution,and at a galactocentric distance <∼3 kpc (see e.g., Fig.
X
it was been suggested that the metal rich group
3 in Zinn 1985). Much progresshas been made in
r ([Fe/H]> −0.8) would constitute a population of
a recent years in the study of this group of clus-
“disk” globular clusters (Zinn 1985; Armandroff
ters, thanks to higher angular resolution CCD
and near-infrared photometry and high resolu-
1This work is based on observations with the
tion spectroscopy. The kinematics, vertical scale
NASA/ESA Hubble Space Telescope, obtained at the
SpaceTelescopeScienceInstitute,operatedbyAURAInc. height, and metallicity distribution of these clus-
undercontracttoNASA tersturnoutto beindistinguishablefromthoseof
2European Southern Observatory, Karl Schwarzschild the bulge field stars (Minniti 1995; Barbuy et al.
Strasse 2, D-85748 Garching bei Mu¨nchen, Germany; 1998, 1999b; Cˆot´e 1999), leaving little doubt that
[email protected],[email protected]
they constituteafamilyofbulge,ratherthandisk
3Universita` di Padova, Dipartimento di Astronomia,
globularclusters(see alsoHarris2000for a recent
Vicolo dell’Osservatorio 5, I-35122 Padova, Italy; or-
[email protected] discussion).
4UniversidadeFederal doRioGrande doSul,Dept. de Among these clusters, the best studied ones
Astronomia, CP 15051, Porto Alegre 91501-970, Brazil;
are the “twin” clusters NGC 6553 and NGC 6528
[email protected]
5Universidade de Sa˜o Paulo, Dept. de Astrono- whichhave virtuallyidentical color-magnitudedi-
mia, CP 3386, Sa˜o Paulo 01060-970, Brazil; bar- agrams (Ortolani et al. 1995). The subject of the
[email protected] present investigation, NGC 6553 (α=18:09:15.6,
1
◦ ◦
δ=-25:54:28, l = 5.25, b = −3.02), is a mod- taken with a time interval of 4.16 years. Both
erately concentrated (central surface brightness: sets were collected with HST using the WFPC2
logΣ = 4.53L⊙pc−2; Djorgovski 1993) globular camera, with the filters F555W (V) and F814W
cluster located at ∼ 3 kpc from the Galactic (I). The first epoch data were acquired by our
Center. Its metallicity is high, although the ex- group on February 1994, as part of the GO5436
act value is still a matter of debate: Barbuy program (Ortolani et al. 1995; Guarnieri et al.
et al. (1999a) and Coelho et al (1999) found 1997). Those of the second epoch were collected
[Fe/H]=−0.5,while Cohenetal. (1999)measured in April 1998, as part of the GO7307 program,
[Fe/H]=−0.16. The first CCD photometry in the and were retrieved from the HST archive. The
BVRI bands showedclear metal-rich characteris- log of observations for the two epochs is shown in
tics,withthepronouncedV bandluminosityturn- Table 1.
over of the upper Red Giant Branch (RGB) to- Since the datasets were taken for independent
wardscoolertemperatures,aneffectofstrongTiO purposes,they covertwo fields which do not com-
blanketing (Ortolani et al. 1990). Subsequent VI pletely overlap. Figure 1 shows the relative orien-
data obtained with the Hubble Space Telescope tationofthemosaics. Blackdotsshowallstarsde-
(HST)byOrtolanietal.(1995;seealsoGuarnieri, tectedintheWFPC2mosaicinthefirstepochob-
Renzini & Ortolani 1997; Guarnieri et al. 1998) servations: the size of the symbol is proportional
reachedmorethan3magnitudesbelowtheturnoff, to the star’s brightness. The cluster center is lo-
leadingtothe conclusionthatNGC 6553isnearly catedapproximatelyatpixel(932,682). Overplot-
coevaltothe haloglobularclusters. Still, thevery tedonthismapisthelocationoftheWFPC2mo-
high contamination by background stars around saic of the second epoch observations, where the
the cluster turnoff was very high, leaving some cluster is centered on WF3. This configuration,
uncertainty on the value of the turnoff luminos- coupled with the different limiting magnitudes of
ity MVTO. The more recent study by Sagar et al. the two observation sets, implies that only about
(1999)overalargerfield(6′×10′)allowedthemto 1/3 of the stars detected in the first epoch data
obtain a cleaner cluster population by statistical have a counterpart in the second epoch photome-
subtraction of background stars. However, their try and could be used for astrometric purposes.
data are not deep enough to allow photometry of
All images were de–biased, dark corrected and
turnoffstarswiththeaccuracyrequiredtoaddress
flatfielded through the standard HST pipeline.
the problem of the cluster age.
Following Silbermann et al. (1996), we have
Using a technique pioneered by King et al. masked out the vignetted region, the saturated
(1998),inthe presentworkwe combineHST wide and bad pixels and columns using a vignetting
field and planetary camera (WFPC2) data from mask, together with the appropriate data quality
two epoch observations of the central regions of file for each frame.
NGC 6553. Thanks to the excellent astrometric
The photometric reduction of eachdatasetwas
performances of WFPC2, we measure the rela-
carriedoutusingtheDAOPHOTII/ALLFRAME
tive proper motion of the cluster with respect to
the bulge. This allows us to i) decontaminate
the CMD of NGC 6553 from the background and TABLE 1
therefore to obtain a more reliable measure of its LOG OFOBSERVATIONS
age;ii)obtainthethreecomponentsofthecluster Program Date Filter Exp. (s)
space velocity; and iii) measure the bulge proper GO5436 25.2.1994 F555W 14
motion dispersions along Galactic longitude and F555W 2×100
latitude. F814W 5
F814W 2×50
2. OBSERVATIONALDATAANDANAL- GO7307 24.4.1998 F555W 3×5
YSIS F555W 3×200
F814W 3×20
We analyse two sets of observations of the cen- F814W 3×200
tralregionofthebulgeglobularclusterNGC6553,
2
package (Stetson 1987, 1994). Preliminary pho- (Holtzman et al. 1995). However, in this case no
tometrywasperformedoneachsingleframeinor- dereddeningcorrectionhasbeenappliedtoourin-
der to obtainan approximatelist ofstars for each strumental magnitudes, because we were mainly
of them. The coordinates of the stars in common interested in producing a CMD as similar as pos-
wereusedforanaccuratespatialmatchamongthe sible to that of the ground-based photometries,
differentframes. Withthe coordinatetransforma- andtoderivethe clusterreddening. Thelatter,in
tions,wethenobtainedasingleimage,whichisthe particular, is found to be strongly variable across
median of all the frames, regardless of the filter. the cluster area(Heitsch & Richtler 1999,see also
In this way we removed all the cosmic rays and Sect. 5) and therefore we would anyway subtract
obtainedthehighestsignal-to-noiseimageforstar an inappropriate reddening to many stars.
finding. The DAOPHOT/FIND routine was ap-
pliedtothemedianimageandPointSpreadFunc- 3. THE COLOR MAGNITUDE DIA-
tion (PSF) fitting photometry was carried out, in GRAM
order to obtain the deepest list of stellar objects,
Figure 2 shows the calibrated CMD extracted
freefromspuriousdetections. Finally,thislistwas
fromthefirstepochdata. Allthestarsdetectedin
used as input to ALLFRAME, for the simultane-
the fourWFPC2 chips areshowninthis diagram.
ousprofilefittingphotometryofindividualframes.
ThemostpopulatedbranchesinthisCMDarethe
The PSFs that we used were the WFPC2 model
cluster Main Sequence (MS), Red Giant Branch
PSFs extracted by P. B. Stetson (1995, private
(RGB), and Horizontal Branch (HB): the latter
communication)fromalargesetofuncrowdedand
beingtheinclinedclumpofstarsatV ∼16.6. The
unsaturated images. It is worth emphasizing that
narrow sequence of stars at V ∼ 17.5, wich is al-
the photometric reductions of the two datasets
mostparalleltotheHBsequenceistheRGBbump
were carried out separately.
(Iben, 1968); as predicted by theory (e.g., Rood
The quality of the first epoch photometry
1972), it is prominent in this metal-rich cluster.
turned out to be better than that of the second
Both the RGB bump and the HB are shaped as
epoch, even if the total exposure time for each
aninclinedsequence(insteadofaclump)bydiffer-
filter was longer in the second epoch data. The
ential reddening acrossthe field (see below). Also
difference in the quality of the photometry in the
clearly visible is the bulge RGB, which is about
two epochs is likely to be due to a combination of
0.5magnitudesredderand∼1.5magfainterthan
twoeffects: the worseningwith time ofthe charge
the cluster RGB (the cluster is ∼ 3 kpc closer to
transfer efficiency loss of the WFPC2 chips, and
usalongthelineofsightthanthebulkofthebulge
the fact the adopted model PSFs, being obtained
field star population; (see below), being the clus-
in 1995, reproduced the actual PSF of the first
ter ∼ 3 kpc closer to us compared with the bulk
epoch better than the second. In what follows
of bulge stars. The bulge Sub Giant Branch and
we will base our analysis on the 1994 photome-
turnoffregionatV ∼21.5and(V −I)∼2.1over-
try, while the spatial information from the 1998
laps with the MS of NGC 6553. A well populated
datacombinedtothatof1994willbeusedforthe
blue sequence departing from the cluster turnoff
relative proper motion derivation.
is also evident, and is expected to be a mixture of
Calibration of the HST instrumental magni-
cluster blue stragglers and foregrounddisk stars.
tudes to the Landolt system was performed fol-
Fortheastrometriccalculationsonlythosestars
lowingtheprescriptionsbyDolphin(2000). These
in common between the two epoch data, having
are an update of the widely used Holtzman et al.
photometric error smaller than 0.07 mag in both
(1995) equations, including the most recent cor-
bands and shape parameter −0.02 <SHARP<
rection for the charge transfer efficiency loss. In
0.05 were selected. Polynomial 20-coefficient co-
principle, since the HST filter passbands are dif-
ordinate transformations were then calculated in
ferent from the Landolt ones, the recommended
order to take into account both the differences in
procedure to transform the instrumental magni-
the pointings and the distortion of the WFPC2
tudes into the standard ground-based system is
field. Figure 3 shows the residuals differences be-
to subtract the absolute extinction in the instru-
tween the coordinates of the two epochs (epoch2-
mental bands before applying the transformations
3
epoch1). Obviously, since the astrometric trans- http://physun.physics.mcmaster.ca/Globular.html
formations were constrained mainly by the much ), and adopting a Sun motion of (U⊙,V⊙,W⊙) =
more numerous cluster stars,the residuals are,by (10, 5.25, 7.17) km s−1 (Dehnen & Binney 1998),
definition, distributed around zero. Foreground the rotational velocity of the Local Standard of
and background (mostly bulge and some disk) Rest of V = 239 km s−1 (Arp 1986) and a
LSR
stars, instead, are distributed around (dx,dy) = cluster distance d=5300pc (see below)the three
(−0.23,0.05)inthisplot. AlsoshowninFig.3are components of the NGC 6553 absolute space ve-
thedirectionsofincreasingGalacticlongitudeand locity are derived: (Π,Θ,W)=(-3.5, 230, -3) km
latitude. s−1 [Π points radially outwardsfrom the Galactic
Figure4showstheresultoftheastrometricde- center,W towardsthe North Galactic PoleandΘ
contamination of the CMD. The left panel shows is oriented in the direction of Galactic rotation].
the CMD of only those stars which displacement This assumesthat we measure allthe background
between the two epochs lies inside a radius of 0.1 bulge stars, with those in the near side moving
pixels centered on (0,0) in Fig. 3, which, as ex- on average in the opposite direction with respect
pected, are mainly cluster stars. Many cluster to those in the far side because of bulge rota-
starsfalloutsidethiscircle,butsuchasmallradius tion. Therefore, their average proper motion is
was chosen in order to select the cleanest sample zero. One concludes that NGC 6553 is describing
ofclusterstars. TherightpanelofFig.4showthe a circular orbit close to the Galactic plane. Its
stars having proper motions dx < −0.15. Those rotationalvelocity is consistentwith the meanro-
are mainly bulge stars, with some residual con- tation of the bulge at 2.7 kpc (Minniti 1995) but
tamination from cluster outliers. also with the disk rotation at the same distance
(e.g. Amaral et al. 1996). Hence the kinematics
4. PROPER MOTIONS AND SPACE of NGC 6553appears to be consistent with either
VELOCITY a disk or a bulge membership. However, its stel-
lar population (age and metallicity) is identical
4.1. NGC 6553 Proper Motion to that of another metal rich cluster, NGC 6528
(c.f., Sec.5), suggestingthat the twoobjects have
ThedistributionoftheresidualsshowninFig.3
formed within the same environment (Ortolani et
allows us to determine both the relative proper
al. 1995). NGC 6528 is located at 1.3 kpc from
motion of NGC 6553 with respect to the bulge,
the Galactic center, and its high radial velocity
and the dispersion of the bulge proper motions
(vr = 185 km/s; Harris 1996) clearly indicates
along the two Galactic coordinates.
that it is on a highly eccentric orbit, therefore ex-
The former is the difference between the cen-
cludingdiskmembershipforthisbulgecluster. All
troids of the bulge and cluster proper motion
in all, both clusters appear to belong to the bulge
distributions, obtained by means of a Gaussian
population of globular clusters, with their kine-
fit along both the l and b axis. Figure 5 shows
maticalpropertiesbeingwithinthe corresponding
the histograms of the dl and db distributions
distribution for bulge stars.
of both cluster and bulge stars. Note that the
ratio of the integrals of the two Gaussians in 4.2. Bulge Proper Motion Dispersion
Fig. 5 (lower left) also gives the relative num-
ber of bulge to cluster stars in the region of the The σ of the distributions shown in Figure 5
WFPC2fieldincommonbetweenthe twoepochs, also allow the determination of the bulge proper
which is 1400:7900=1:5.6. The relative motion of motiondispersion. Inordertoreducethecontam-
the cluster stars is by construction centered on ination due to cluster stars in the determination
dl = 0 and db = 0, while the bulge stars are clus- of the bulge velocity dispersion σl along the axis
tered around (dl,db) = (−0.245,−0.017) pixels, of Galactic longitude, bona fide bulge stars were
which corresponds to a relative proper motion of selectedfromtheCMDasindicatedinFig.5. The
NGC 6553 with respect to the bulge of µl = 5.89 distributionofthesestarsinthe(dl,db)plane(up-
and µb = 0.42 mas yr−1. Combining this result per right panel) gives indeed a cleaner sample of
with the radial velocity of NGC 6553, vr = −6.5 bulge stars, with some residual cluster members,
(Harris 1996, as updated on the WEB page: comingfromtheclusterlowermainsequence. The
4
Gaussianfittothisdistributionshowninthelower presents lower reddening, both absolute and dif-
panels yields a better determination of the bulge ferential, and its CMD has much narrower and
proper motion dispersion: σl =0.119±0.012 and bluer sequences. A fiducial ridge line extracted
σb = 0.098±0.009. The quoted uncertainties are from it was used to apply the method described
the formal errors in the Gaussian fit of the his- in Piotto et al. (1999) and obtain a cluster CMD
togram shown in Fig. 5. The fit was performed partially corrected for differential reddening. The
takingintoaccounterrorsinbothcoordinates: the observedfieldwasdividedin84sub-fieldsofequal
Poisson error for the y and the σ of the cluster size and the CMD of each of them was shifted
Gaussian for the x. The dotted lines in the lower alongthe reddeningline tooverlapthe fiducialse-
left panel are the two Gaussians whose sum is the quence. TheresultisshowinFig.7: theleftpanel
solidline. Theresidualclustercomponentwasfit- showsthe originalclusterCMD,the centraloneis
ted with a Gaussian with the same σ obtained reddening-correctedandtherightpanelshowsthe
from the fit of Fig. 5. Formally, the space veloc- fiducialline for47Tuc andNGC6528overplotted
ity dispersion corresponding to the measuredσ of on that of NGC 6553.
the cluster stars is 28 km s−1. We are not aware The central and right panels of Fig. 7 allow a
of any measurements of the velocity dispersion of measure of age and distance with higher accuracy
NGC 6553, but in general globular clusters have withrespecttothe previouslyavailabledata. The
velocity dispersions < 10 km s−1 (e.g., Pryor & estimated apparent magnitude of the cluster HB
Meylan 1993). Therefore, we conclude that the is VHB = 16.6±0.05. Following Ortolani et al.
distributionoftheclusterstarinthe(dl,db)plane (2000),we adopt MHB =(0.16±0.10)×[Fe/H]+
V
isentirelydue to observationalerrors. Deconvolv- (0.98±0.10)and[Fe/H]=−0.5±0.3,givingMHB =
V
ing the latter from the observed dispersion of the 0.90±0.12. Theclusterapparentdistancemodulus
bulge stars we finally obtain σl =2.63±0.29 and is therefore (m−M)V = 16.60−0.90 = 15.70±
σb = 2.06 ± 0.21 mas yr−1. We therefore con- 0.13.
firm the existence of a proper motion anisotropy
The reddening is derived from comparisons
σl/σb = 1.28±0.19 already found by many au- to the 47 Tucanae locus (Ortolani et al. 1995;
thors (e.g., Zhao, Rich & Biello, 1996, who mea-
Kaluzny et al. 1997). The color of the RGB
sured σl/σb = 1.15±0.06 towards Baade’s Win- of NGC 6553 at the HB level is V − I =
dow) and interpreted as evidence of the mean 2.08±0.03. With respect to 47 Tuc we de-
bulge rotation. The values of the proper motion rive ∆(V − I) = 0.99. Part of
(NGC6553−47Tuc)
dispers◦ions obtaine◦d from Fig. 6 for our field at this color shift (∆(V − I) = 0.134, Girardi
(l = 5.25,b = −3.02) are somewhat lower than
et al. 2000) is due to the metallicity differ-
the values σl ∼ 3.2 and σb ∼ 2.7 found both to- ence between the two clusters. Indeed, 47 Tuc
wards Baade’s Window (l =1◦,b= −4◦) and the
has metallicity [Fe/H]=–0.76 (Harris 1996) and
Plaut’sField(l =0◦,b=−8◦)(Spaenhauer,Jones
[α/Fe]=0.3 (Gratton, Quarta & Ortolani 1986),
& Whitford1992;Rich&Terndrup1997;Mendez
therefore a global metallicity [M/H]=–0.55, while
et al. 1997).
Coelho et al. (2000) measured [M/H]=-0.1 for
NGC 6553. Assuming E(V − I) = 0.05
47Tuc
5. THEDISTANCEANDAGEOFNGC6553 (Barbuy et al. 1998) we finally obtain E(V −
I) = 0.99 − 0.135 − 0.05 = 0.805 and
The relative proper motion decontamination NGC6553
therefore E(B−V) = E(V −I)/1.28 = 0.63 and
has provided a cleaner cluster CMD, but the se-
A = 0.63 × 3.3 = 2.08. The adopted total-
quences still show a significant width due to dif- V
to-selective extinction ratio R = 3.3 ratio was
ferentialreddening(e.g. Heitsch&Richtler1999). V
assumed considering the metallicity and redden-
Intheprocessofmatchingtheoverlappingregions
ing amount dependences (Barbuy et al. 1998 and
of different frames in the two epochs data, the
references therein). This value of the reddening
WFPC2 frame was divided into small sub-fields
is lower than previous determinations (and there-
(e.g., part of the first epoch WF4 overlaps partly
fore the apparent distance modulus shorter) be-
withthe WF3 andpartlywith WF4 ofthe second
causethe reddeningcorrectedCMDhas beenreg-
epoch data). In these comparisons, it appeared
istered on the fiducial line of a low-reddening re-
that a region of the cluster mapped by the WF4
5
gion. TheabsolutedistancemodulusofNGC6553 tions. This is displayed in the lower right panel
is (m−M) = 15.70−2.08 = 13.62, in perfect of Fig. 8 which shows all the stars brighter than
0
agreementwiththevalue(m−M) =13.60found V = 19.3 and bluer than V − I = 1.76 in the
0
by Guarnieriet al.(1997),andcorrespondingtoa original CMD (Fig. 2), and confirms that a large
heliocentric distance d=5.3 kpc. fraction of them are indeed concentrated around
Figure 7 also shows the fiducial line of 47 Tuc (dx,dy)=(0,0) and therefore are likely BS mem-
and NGC 6528 (Ortolani et al. 1995) overplot- bers of NGC 6553.
ted on that of NGC 6553. The former two were
shifted bothinmagnitude andin colorin orderto 7. CONCLUSIONS
bring their upper MS into coincidence with that
Using WFPC2 images taken 4.16 yr apart we
of NGC 6553. The extremely small differences in
have been able to measure both the proper mo-
the TO and HB locations are well inside the un-
tion of the globular cluster NGC 6553 and of the
certainty in the construction of the fiducial lines,
Galacticbulgestarsinthebackgroundoftheclus-
fromwhichweconcludethatthethreeclustersare
coeval to within ∼1 Gyr, and therefore the abso- ter. Themainresultsobtainedcanbesummarized
as follows.
luteageofNGC6553(andNGC6528)is∼13Gyr
1. The cluster NGC 6553 appears to be on a
as recently determined for 47 Tuc (Zoccali et al.
nearly circular orbit ∼ 3 kpc from the center of
2001). This demonstrates that the Galactic bulge
the Galaxy, with a velocity of ∼ 230 km s−1 and
wasenrichedtosolarabundancetowithinarather
a small inclination with respect to the Galactic
shorttime scale(less thana few Gyr). The differ-
plane. Thesekinematicalpropertiesareconsistent
ence in the RGB slopes seen in the right panel of
with both a bulge and a disk membership of the
Fig.7 is due to the difference in globalmetallicity
cluster. However, the cluster stellar population
of the three clusters. Note that the age-sensitive
parameter∆VTO changesbyonlyafewhudredths is virtually identical to that of the cluster NGC
HB
6528, which kinematics makes its membership to
of magnitudes in this metallicity range (see Fig. 3
thebulgeunambiguous,andweconcludethatalso
of Rosenberg et al. 1999).
NGC6553islikelyamemberoftheGalacticbulge.
6. THECLUSTERBLUESTRAGGLERS 2. Therelativeastrometrybetweenthetwoepochs
isaccurateenoughformeasuringthedispersionof
As apparent from Fig. 4, at least part of the propermotionsofbulgestarsalongthetwoGalac-
stars in the blue sequence above the cluster MS tic coordinates, with σl =2.63 and σb =2.06 mas
seemtobelongtothecluster,hencetheyareprob- yr−1.
ably blue stragglers (BS), rather than Galactic
3. The distinct proper motions of the cluster and
disk main sequence stars. This is confirmed by
bulge starshas allowedus to constructdecontam-
the following two experiments.
inatedCMDsforboththeclusterandbulgepopu-
In order to check this issue, we isolated the lations. TheresultingimprovedCMDfortheclus-
probable BS stars in the decontaminated CMD ter has allowed us to obtain more accurate values
(Fig. 8: left panel), and investigated their radial forthereddening,distance,andageofthecluster:
distribution as compared with the other cluster NGC6553appearstohavethesameageasthein-
stars. The upper right panel of Fig. 8 shows that nerhaloglobularcluster47Tuc(∼13Gyr)within
thecumulativedistributionoftheBS(dottedline) a ∼1 Gyr uncertainty. Giventhe highmetallicity
have the same spatial distribution as the cluster of both NGC 6553 and NGC 6528, this indicates
RGB (solidline)with 17.5<V <19.6. This indi- thattheGalacticbulgeunderwentrapidmetalen-
catesthat they areindeedclustermembers. How- richmenttosolarmetallicity andbeyond,some13
ever, the BS population of NGC 6553 does not Gyrago,supportingtheviewofanearlyandrapid
appear concentrated towards the center as they assembly of the Galactic Bulge.
are in high density clusters (e.g., Ferraro et al.
4. Thanks to the accurate astrometry obtained,
1999). The second check on the BS nature of a
we have also demonstrated that the cluster con-
major fraction of the blue sequence stars can be
tains a significant population of blue stragglers,
made by using their distribution of proper mo-
having separatedthem from a trace population of
6
starswithsimilarluminositiesandcolorswhichin- Heitsch, F., & Richtler, T. 1999, A&A, 347, 455
stead are likely intermediate age foreground disk
Holtzman J., Burrows C.J., Casertano S. et al.,
stars.
1995,PASP 107, 1065
Iben, I. Jr. 1968, ApJ, 154, 581
We thank the referee, Patrick Cˆot´e, for useful
comments. BBandEBacknowledgepartialfinan- Kaluzny,J.,Krzeminski,W.,Mazur,B.,Wysocka,
cial support from CNPq and Fapesp (Brasil). SO A., Stepien, K. 1997, Acta Astron., 47, 249
acknowledgesthe Ministero dell’Universit`ae della King I.R., Anderson J., Cool A.M., Piotto G.,
RicercaScientifica e Tecnologica(MURST) under 1998,ApJ, 492, L37
the program on ’Stellar Evolution’ (Italy).
Mendez, R.A., Rich,R.M., VanAltena, W.F., Gi-
rard, T.M., van den Berg, S., & Majewski, S.,
REFERENCES
1998,IAU Symp.
Amaral, L.H., Ortiz, R., L´epine, J.R.D., Maciel, Minniti, D., 1995, AJ, 109, 1663
W.J. 1996, MNRAS, 281, 339
Ortolani S., Barbuy B., Bica E. 1990, A&A 236,
Armandroff, T.E. 1989, AJ 97, 375 362
Arp, H. 1986,A&A, 156, 207
OrtolaniS., Renzini A., Gilmozzi R., Marconi G.,
Barbuy B., Bica E., Ortolani S. 1998, A&A, 333, Barbuy B., Bica E., Rich R.M. 1995, Nature
117 377, 701
Barbuy, B., Renzini, A., Ortolani, S., Bica, E., & Ortolani,S.,Momany,Y.,Bica,E.,&Barbuy,B.,
Guarnieri, M.D. 1999a,A&A, 341, 539 2000,A&A, 357, 495
Barbuy B., Ortolani S., Bica E., Desidera S. Piotto,G.,ZoccaliM.,KingI.R.,DjorgovskiS.G.,
1999b, A&A 348, 783 SosinC.,RichR.M.,MeylanG.,1999,AJ,118,
1737
Coelho, P., Barbuy, B., Perrin, M.N., Idiart, T.,
& Schiavon, R.P., 2000, A&A, in preparation Pryor,C.,&,Meylan,G.,1993,inASPConf.Ser.50,
Structure and Dynamics of Globular Clusters,
Cˆot´e, P., 1999, AJ, 118, 406
eds. S.G. Djorgovski and G. Meylan, (San
Dehnen, W., & Binney, J.J. 1998, MNRAS, 298,
Francisco: ASP), 357
387
Rich, R.M., & Terndrup, D.M., 1997, in ASP
Djorgovski,S.1993,inStructureandDynamicsof
Conf.Ser. 127, Proper Motions and Galactic
Globular Clusters, eds. S. G. Djorgovski & G.
Astronomy, ed. R.M. Humphreys, (San Fran-
Meylan (San Francisco: ASP), p. 373
cisco: ASP), 129
Dolphin A.E., 2000, PASP, 112, 1397
Rood, R.T., 1972,ApJ, 177, 681
Ferraro, F.R., Paltrinieri, B., Rood, R.T., Dor-
Rosenberg,A.,Saviane,I.,Piotto,G.,&Aparicio,
man, B., 1999, ApJ, 522, 983
A., 1999,AJ, 118, 2306
Girardi,L.,Bressan,A., Bertelli,G., & Chiosi,C.
Sagar, R., Subramaniam, A., Richtler, R., and
2000,ApJ, 530, 62
Grebel, E.K. 1999, A&A, 135, 391
Gratton, R.G., Quarta M.L., & Ortolani S. 1986,
Silbermann N.A., Harding P., Madore B.F., 1996,
A&A, 169, 208
ApJ 470, 1
Guarnieri,M.D.,Renzini,A.,&Ortolani,S.1997,
Spaenhauer, A., Jones, B.F., & Whitford, A.E.,
ApJ, 477, 21
1992,AJ, 103, 297
Guarnieri M.D., OrtolaniS., Montegriffo P., Ren-
Stetson P.B., 1987, PASP 99, 191
zini A., Barbuy B., Bica E., Moneti A., 1998,
A&A 331, 70 Stetson P.B, 1994, PASP 106, 250
Harris W.E. 1996, AJ 112, 1487 Zhao,H.,Rich,R.M.,&Biello,J.,1996,ApJ,470,
506
Harris,W.E.2000,SaasFeeLectures1998(Berlin,
Springer), in press Zinn, R., 1985,ApJ, 293, 424
7
Zoccali, M., Renzini, A., Ortolani, S., Bragaglia,
A., Bohlin, R., Carretta, E., Ferraro, F.R.,
Gilmozzi,R.,Holberg,J.B.,Marconi,G.,Rich,
R.M., Wesemael, F., 2001, ApJ, submitted
This2-columnpreprintwaspreparedwiththeAASLATEX
macrosv5.0.
8
Fig. 1.—X,YpixelmapofNGC6553: allstarsdetectedintheWFPC2mosaicinthefirstepochobservations
are represented by black dots; symbol sizes are proportional to the stellar brightness. The location of the
WFPC2 mosaic of the second epoch observations is overplotted on this map.
9
Fig. 2.— NGC 6553 calibrated CMD extracted from the four WFPC2 chips of the first epoch observations.
10