Discovery of more than two thousand white dwarfs in the globular cluster Omega Centauri

0.45 MB·English

by  M. Monelli

Draftversion February2,2008
PreprinttypesetusingLATEXstyleemulateapjv.6/22/04
DISCOVERY OF MORE THAN TWO THOUSAND WHITE DWARFS IN THE GLOBULAR CLUSTER
ωCENTAURI1
M. Monelli 2, C.E. Corsi 2, V. Castellani 2,3, I. Ferraro 2, G. Iannicola 2, P.G. Prada Moroni 4,5,6, G. Bono 2,
R. Buonanno 7, A. Calamida 2,7, L.M. Freyhammer 8,9, L. Pulone 2, and P.B. Stetson 10
(Dated: drafted February 2, 2008 / Received / Accepted)
Draft versionFebruary 2, 2008
ABSTRACT
Wepresentdeepmultiband(F435W,F625W,andF658N)photometricdataoftheGlobularCluster
ω Cen collected with the Advanced Camera for Surveys on board of the Hubble Space Telescope.
5 We identified in the (F435W-F625W, F435W) plane more than two thousand White Dwarf (WD)
0 candidatesusingthreeoutofnineavailablepointings. Suchalargesampleappearsinagreementwith
0
predictions based on the ratio between WD and Horizontal Branch (HB) evolutionary lifetimes. We
2
alsodetected≈1600WDsinthe(F658N-F625W,F625W)plane,supportingtheevidencethatalarge

n fraction of current cluster WDs are Hα bright.
a Subject headings: globular clusters: general — globular clusters: omega Centauri
J

8
2 1. INTRODUCTION He content among MS stars.
Even though ω Cen presents properties which need to
Among Galactic Globular Clusters (GGCs), ω Cen is
v1 the most massive one (M = 5 × 106M⊙, Meylan et btoepinrvoepsetriglyatuendseevrestroaoldo,pietnsspterollbalrecmosntceonntciesranignogldstmelilnaer
al. 1995) and the only one which clearly shows a well-
3 evolution and its dependence on the metallicity. This
defined spread in metallicity. According to recent es-
3 applies not only to evolved stars such as RR Lyrae, hot
timates based on sizable samples of evolved Red Giant
6
HB stars, and the tip of the RGB, but also to the ex-
1 Branch (RGB) and sub-giant branch stars, the metallic-
0 itydistributionshowsthreepeaksaround[Fe/H]=−1.7, pected population(s) of WDs. The search for WDs in
5 −1.5and−1.2togetherwithatailofmetal-richstarsap- GGCshasalreadybeensuccessful,andseveralWDsam-
0 proaching[Fe/H]≈−0.5(Norrisetal.1996;Hilkeretal. pleshavebeenidentified(Hansenetal.2003;Moehleret
al.2004). Sofar,therichestsampleofclusterWDs(222)
/ 2004). During the last few years it has also been sug-
h was detected in M4 by Hansen et al. (2004). The detec-
gested that ω Cen harbors multiple stellar populations
p tion of WDs in ω Cen dates back to Ortolani & Rosino
(Lee et al. 1999)characterizedby different ages (Ferraro
- (1987) who selected two dozen WD candidates on the
o etal.2004;Hughesetal.2004),Heabundances,anddis-
r tances (Bedin et al. 2004; Norris 2004; Freyhammer et basis of ground-based data, and to Elson et al. (1995),
t who detected four WD candidates using HST data. In
s al. 2005). More recently, Sollima et al. (2005)suggested
a thatRGBstarspresentfivepeaksduetodifferentstellar thispaper,wepresentpreliminaryresultsconcerningthe
: identification of more than 2,200 WDs in ω Cen on the
v populations. Moreover, Piotto et al. (2005) found, that
i the bluer main sequence (MS) detected by Bedin et al. basisofB,R,HαdatacollectedwiththeAdvancedCam-
X era for Surveys (ACS) on board of HST, and available
(2004) is more metal-rich than the red MS. To account
on the HST archive. The reader interested in thorough
r forthisfeaturetheysuggestedapossiblevariationofthe
a reviews concerning recent theoretical and empirical re-
1 Based on observations collected with the Advanced Camera sultsisreferredto(Fontaine,Brassard,&Bergeron2001;
forSurveysonboardoftheHubbleSpaceTelescope. Koester 2002; Prada Moroni & Straniero 2002; Hansen
2 INAF-Osservatorio Astronomico di Roma, via Frascati 33, & Liebert 2003; Hansen 2004,and references therein).
Monte Porzio Catone, Rome, Italy; [email protected],
[email protected], [email protected], giac- 2. OBSERVATIONSANDDATAREDUCTION
[email protected], [email protected], pu-
Current data were collected with nine pointings of
[email protected]
3 INFN, Sezione di Ferrara, via Paradiso 12, 44100 Ferrara, the ACS camera across the center of the cluster. The
Italy 3×3 mosaic covers a field of view of ≈ 9′ ×9′. Four
4 Dipartimento di Fisica ”E. Fermi”, Univ. Pisa, Largo B.
images per field have been acquired in three different
Pontecorvo2,56127Pisa,Italy;[email protected]
5 INFN,Sez. Pisa,viaE.Fermi2,56127Pisa,Italy bands, namely F435W, F625W, and F658N (hereinafter
6 INAF-Osservatorio Astronomico di Teramo, via M. Maggini, B, R, and Hα). Three deep (340s) exposures were
64100Teramo,Italy secured for the B and R-band, respectively, while the
7 Universita` di Roma Tor Vergata, via della Ricerca Sci-
entifica 1, 00133 Rome, Italy; [email protected], exposure time for the four Hα images was 440s each.
[email protected] The nine fields were independently reduced with the
8 Royal Observatory of Belgium, Ringlaan 3, B-1180 Brussels, DAOPHOTII/ALLFRAME (Stetson 1994). An individ-
Belgium ual PSF has been extracted for each frame by adopting,
9 Vrije Universiteit Brussel, OBSS/WE, Pleinlaan 2, B-1050
on average, ≈200 bright isolated stars. The individual
Brussels,Belgium;[email protected]
10 Dominion Astrophysical Observatory, Herzberg Insti- catalogues were rescaled to a common geometrical sys-
tute of Astrophysics, National Research Council, 5071 West tem with DAOMATCH/DAOMASTER. The final cat-
Saanich Road, Victoria, British Columbia V9E 2E7, Canada; alogue includes approximately 1.2×106 stars, and the
[email protected]
B,B −R CMD shows a well-populated cluster MS to-
2 Monelli et al.
gether with a sizable sample of stars covering a wide at least below MB ∼7. According to predictions by Al-
region to the left of the MS (Bono et al. 2005). thaus, & Benvenuto (1998) for a 0.5M⊙ WD with a CO
From this catalogue we selected all the stars bluer core and pure H atmosphere models by Bergeron, We-
than MS stars and fainter than extreme HB stars. We semael, & Beauchamp (1995)12 the current WD sample
ended up with a sample of approximately 45,000 stars provides the opportunity to investigate the WD cooling
distributed over the nine pointings. Among them we se- for luminosities ranging from ten times the solar lumi-
lectedthreepointings(2,7,and9)thatinclude≈14,000 nosity down to logL/L⊙ ∼−3.1 (B ∼27, MB ∼12.8).
WD candidates. These stars have been identified in in- By adopting the same theoretical predictions one can
dividual deep B,R,Hα images and we performed once easily recognize that the huge number of WDs should
again the photometry using ROMAFOTwo. Individual not be a real surprise. As a matter of fact, the num-
starshavebeeninteractivelycheckedineveryimage,and ber ratio between WD and HB stars, for not too long
theWDcandidatesweremeasuredeitherasisolatedstars cooling times, is simply given by the ratio of the life-
or together with neighbor stars, where crowding affects times spent during these two evolutionary phases. On
the photometric accuracy. Note that a substantial frac- the basis of the shallow ACS photometry in the B,R
tionofthestarslocatedclosetoWDcandidatesaretruly bands, recently presented by Freyhammer et al. (2005),
MS stars, i.e. they did not belong to the original sam- wefoundinthesamethreefieldstheoccurrenceof∼630
ple of stars located on the blue side of the MS. Note HB stars. The typicalevolutionary lifetime for HB stars
that the 80% of the selected stars revealed to be either is ∼ 1−2×108 year, depending on their ZAHB effec-
cosmic rays, or spurious identifications of faint stars lo- tive temperature (Castellani et al. 2004), while for a
cated close to saturated stars, or too faint objects to WD with a CO core and M = 0.5M⊙ at MB ∼ 11.3
be safely measured on individual images. The photo- (logL/L⊙ = −2.24) the lifetime is ∼ 3.5× 108 years.
metric calibration was performed in the Vega System This means that we expect to detect roughly twice as
(http://www.stsci.edu/hst/acs/documents). manyWDsthanHBstarsbrighterthanB ≈25.5. Inter-
estingly enough, we detected approximately 1200 WDs
brighterthanthismagnitudelimit. However,thecurrent
3. RESULTSANDDISCUSSION
estimate should be considered as a robust lower limit,
Inthisinvestigationwepresentpreliminaryresultsbased sincepreliminary artificialstarexperiments suggestthat
on three out of the nine pointings. Figure 1 shows the at this limiting magnitude the completeness is ≈85%.
currentsampleofω Centauri starsintheB−R,Bplane. AnunexpectedobservedfeatureinFig. 1,isthesteady
Interestingly enough, the original sample splits into two increase in color dispersion when moving toward fainter
well-definedsequences,theredderonemadebyMSstars WD magnitudes. In order to assess whether this spread
andtheotheronemadeby2,212blueobjectsthatranges in color might be due to photometric errors, we per-
from B ≈22, B−R≈0, down to B ≈27, B−R≈0.8, formed an empirical test. We estimated the ridge line
covering the expected region of cluster WDs. of WDs and the distance in color of individual objects
Using GOODS data collected with ACS in the bands from the ridge line. Figure 2 shows the color distribu-
F435W and F606W, we estimated that the expected tion of WDs in the magnitude interval B = 25 ± 0.5
number of field galaxies with 22 ≤ B ≤ 26 and - together with its gaussian fit (solid line). Then we esti-
0.2 ≤ B −R ≤ 0.511 in the same area covered by cur- mated standard deviation in color of the same selected
rent observations is ≈ 60 (Grazian et al. 2005, priv. WDs(intrinsic errors)andwefoundσB−R =0.087. The
comm.). The number of field stars is also negligible, be- dashed line plotted in Fig. 2 shows the expected color
cause halo and disk stars peak around B−R=0.7 and distributionforthesamesampleofWDs(712)according
B − R = 1.8, respectively (King et al. 1990). More totheassumptionthattheirB,Rmagnitudeswouldonly
detailed estimates based on radial velocity (Suntzeff & be affected by gaussian intrinsic errors. Data plotted in
Kraft 1996)or proper motion (van Leeuwen et al. 2000) Fig. 2 indicate that the sigma of the gaussian fit to ob-
measurements in ω Cen suggest that at most two dozen served WDs is a factor of two larger than expected. On
of field stars might be located inside the area covered this basis, we estimated that the two distributions dif-
by current data. Finally, theoretical Galactic models fer at 99% confidence level. This finding, taken at face
(Castellani et al. 2002) suggest that in the same area value, indicates that the color dispersion might be real.
covered by current observations are present at most six- InordertoinvestigatetheWDlocationintheCMdia-
teenfieldWDsbrighterthanB =25.5andfifty-fivefield gramwe adoptedthe WD coolingsequencesforCO core
WDs brighter than B =28. and H envelopes constructed by Althaus, & Benvenuto
This evidence suggests that we are facing with a bona (1998). Theoretical predictions have been transformed
fide sample of cluster WDs, including more than 2,000 intotheobservationalplanebyusingpureHatmosphere
objects, thus the largest sample of WDs ever observed models constructed by Bergeron et al. (1995). Pre-
in a stellar cluster. Data plotted in Fig. 1 clearly show dicted cooling sequences for M = 0.5,0.7,0.9M⊙ have
that, thanks to the sizable sample of ω Cen stars, the been plotted by adopting canonical estimates for cluster
cooling sequence show up at B ≈ 21. This bright limit, reddening (E(B−V)=0.11±0.02, Lub 2002) and dis-
if we assume for ω Cen an apparent distance modulus tance modulus (µ=13.7±0.11,Thompsonet al. 2001).
DMB ∼14.21(Thompsonetal. 2001),implies thatcur- The reddening in the B,R,Hα bands was estimated us-
rent data are tracing the cooling history of cluster WDs ing the extinction model of Cardelli et al. (1989).
The comparisonbetween theory and observations (see
11 Note that to perform this estimate we accounted for the panel a) of Fig. 3) discloses a good agreement for the
difference in magnitude between AB and Vega systems and for
the difference between the filter F606W and the filter F625W
(http://www.stsci.edu/hst/acs/documents). 12Seealsohttp://www.ASTRO.UMontreal.CA∼bergeron/CoolingModels/
More than two thousand WDs in ωCen 3
bright portion (22 ≤ B ≤ 23). However, the theoretical bution. This working hypothesis, once confirmed, indi-
sequences toward fainter magnitudes appear to fit the cates that a substantial fractionof WDs in ω Cen might
blue (hot) edge of the observed WDs. It is noteworthy, be the evolutionary aftermath of binary systems. Ac-
that a WD with a stellar mass M = 0.5M⊙ is a lower cording to current knowledge, candidate cluster He core
limit for WDs with CO-core, actual WDs are expected WDs have only been identified in systems with high
to be slightly more massive (∼ 0.53M⊙, Renzini et al. central densities and short two body relaxation times
1996). Current uncertainties on reddening (0.02 across (NGC6397, logρ0 = 5.68 L⊙pc−3, Taylor et al. 2001;
the cluster, Schlegel et al. 1998) and on cluster distance 47 Tuc, logρ0 = 4.77 L⊙pc−3, Edmonds et al. 2001).
cannotaccountfortheobservedsystematicdriftincolor. However,ω Cen presents a relativelylow centraldensity
This evidence suggests that the observed WD cooling (logρ0 = 3.12 L⊙pc−3), and therefore fills an empty re-
sequence is redder than expected. gion of the parameter space predicted by Hansen et al.
In principle one can find several plausible reasons for (2003,seetheirfigure8). Currentfindingsaremarginally
such an occurrence. In particular, we note that for affected by the adopted theoretical predictions. Differ-
22 ≤ B we are already below the so-called ”DB gap” ent sets of WD cooling sequences (Fontaine et al. 2001;
(Hansen&Liebert2003),i.e. WDswithHeatmospheres Prada Moroni & Straniero 2002), transformed into the
couldbepresent. Dataplottedinpanelb)ofFig. 3,show B−R,B plane by adopting the same atmosphere mod-
thatbyadoptingWDcoolingsequencesforCOcore,and els, agree quite well with each other.
He envelopes by Benvenuto, & Althaus (1997) together Finally,itisworthnotingthatcurrentHα−R,Rdata
with He atmosphere models by Bergeron et al. (1995), (see Fig. 4), show that a good fraction of detected WDs
thecoolingsequenceisindeedmovingtowardreddercol- (∼ 1,600)areHα bright, andindeed they attain Hα−R
ors. Note thatthe cut-offofthe coolingsequences inthe colors systematically bluer than predicted. We would
brightregionisduetothefactthatHeatmospheremod- like to mention that blending and/or binarity could also
elsdonotcoverthistemperatureregion. Thecomparison affectobservedcolors. Therefore,no firmconclusioncan
between theory and observations indicates that the bulk be reached on the basis of current data. More detailed
of the observed WDs might be of the DB type. This photometricandspectroscopicinvestigationsofthislarge
evidence, as suggestedby the referee, is in contrast with sample of cluster WDs will provide fundamental hints
current spectroscopic findings by Moehler et al. (2004) concerning their cooling and progenitors.
concerning cluster WDs. They identified of H Balmer
lines in four and seven WDs in NGC 6397 and in NGC
6752, thus suggesting that these objects are of the DA 4. ACKNOWLEDGMENTS
type. The same outcome applies to field WDs for which It is a real pleasure to thank P. Bergeron not only
DA WDs are ≈ 80% of the entire sample (Koester & for sending us bolometric corrections and color indices
Chanmugam 1990). An explanation of this discrepancy for ACS bands, but also for many useful suggestions.
should await for new spectroscopic measurements. Wealsothank M.CignoniandS.Degl’Innocentiforpre-
As shown in the bottom panel of Fig. 3, a different dictions concerning Galactic models and isochrones. It
possibility is given by the occurrence of He-core WDs. is also with unusual pleasure that we acknowledge an
Interestingly enough, predicted WD cooling sequences anonymousrefereeforhis/herpositivecommentsandin-
for He core structures by Serenelli et al. (2002), trans- sights. This work was supported by PRIN 2003 and by
formed into the observational plane by adopting H at- BFSP within the projects: “Continuity and Discontinu-
mosphere models, also account for the observed distri- ity in the Galaxy Formation“ and “IAP P5/36”.
REFERENCES
Althaus,L.G.,&Benvenuto O.G.1998,MNRAS,296,206 Hansen,B.M.S.,Kalogera,V.,Rasio,F.A.2003,ApJ,586,1364
Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1997, Hansen,B.M.S.,etal.2004,ApJS,accepted, astro-ph/0401443
A&A,327,1054 Hilker,M.,Kayser,A.,Richtler,T.,Willemsen,P.2004,A&A,422,
Bedin,L.R.,etal.2004,ApJ,605,125 9
Benvenuto, O.G.,&AlthausL.G.1997,MNRAS,288,1004 Hughes, J., Wallerstein, G., van Leeuwen, F., & Hilker, M. 2004,
Bergeron, P., Wesemael, F., & Beauchamp, A. 1995, PASP, 107, AJ,127,980
1047 King, I., Gilmore, G., van der Kruit, P. C. 1990, The Milky Way
Bono,G.,etal.2005,in14thEuropeanWorkshoponWDs,ed.D. AsGalaxy, UniversityScienceBooks
Koester,S.Moehler(SanFrancisco:ASP),inpress Koester,D.2002, A&ARv,11,33
Cardelli,J.A.,Clayton,G.C.,Mathis,J.S.1989,ApJ,345, 245 Koester,D.,&Chanmugam, G.1990,RPPh,53,837
Cariulo, P., Degl’Innocenti S., & Castellani, V. 2004, A&A 421, Lee,Y.-W.,Joo,J.-M.,Sohn,Y.-J.,etal.1999,Nature,402,55
1121 Lub, J. 2002, in Omega Centauri: A Unique Window into
Castellani, V.,Cignoni, M.,Degl’Innocenti, S.,Petroni, S., Prada Astrophysics,ed.F.vanLeeuwen,J.HughesandG.Piotto(San
Moroni,P.G.2002,MNRAS,334,69 Francisco:ASP),95
Castellani, V., Zoccali, M., Bono, G., Iannicola, G., Cassisi, S., Meylan, G., Mayor, M., Duquennoy, A., Dubath, P. 1995, A&A,
Buonanno, R.2004,A&A,accepted 303,761
Edmonds,P.D.,Gilliland,R.L.,Heinke,C.O.,Grindlay,J.E.,& Moehler, S., Koester, D., Zoccali, M., Ferraro, F. R., Heber, U.,
Camilo,F.2001, ApJ,557,L57 Napiwotzki,R.,&Renzini,A.2004,A&A,420,515
Elson, R. A. W., Gilmore, G. F., Santiago, B. X., Casertano, S. Norris,J.E.2004, ApJ,612,25
1995, AJ,110,682 Norris,J.E.,Freeman,K.C.,&Mighell,K.J.1996,ApJ,462,241
Ferraro,F.R.,Sollima,A.,Pancino,E.,Bellazzini,M.,etal.2004, Ortolani,S.,&Rosino,L.1987, A&A,185,102O
ApJ,603,81 Piotto,G.,etal.2005,ApJ,accepted, astro-ph/0412016
Fontaine, G.,Brassard,P.,Bergeron,P.2001,PASP,113,409 PradaMoroni,P.&Straniero,O.2002,ApJ,581,585
Freyhammer,L.,Monelli,M.,Bono,G.,etal.2005,ApJ,accepted Renzini,A.,etal.1996,ApJ,465,L23
Hansen,B.M.S.2004,PhR,399,1 Schlegel,D.J.,Finkbeiner,D.P.,&Davis,M.1998,ApJ,500,525
Hansen,B.M.S.,&Liebert,J.2003,ARA&A,41,465
4 Monelli et al.
SerenelliA.M.,AlthausL.G.,RohrmannR.D.,&Benvenuto,O. Thompson,I.B.,etal.2001,AJ,121,3089
G.2002,MNRAS,337,1091 van Leeuwen, F., Le Poole, R. S., Reijns, R. A., Freeman, K. C.,
Sollima, A., Ferraro, F. R., Pancino, E., & Bellazzini, M. 2005, deZeeuw,P.T.2000,A&A,360,472
MNRAS,accepted, astro-ph/0411546
Stetson, P.B.1994,PASP,106,250
Suntzeff, N.B.,&Kraft,R.P.1996,AJ,111,1913
Taylor,J.M.,Grindlay,J.E.,Edmonds,P.D.,&Cool,A.M.2001,
ApJ,553,L169
More than two thousand WDs in ωCen 5
Fig. 1.— Color-Magnitude diagram in the B,B−R bands. The number of WD candidates are labeled. Error bars display intrinsic
photometricerrors.
6 Monelli et al.
Fig. 2.— Color distribution of WDs in the magnitude interval B = 25±0.5. The solid line shows the gaussian fit to the observed
distribution,whilethedashedlinetheexpecteddistributionforthesamesampleofWDsinthecasethattheircolorswouldonlybeaffected
bygaussianphotometricintrinsicerrors.
More than two thousand WDs in ωCen 7
Fig. 3.— Color-Magnitude diagram in the B,B−R bands. Panel a) - Solid, dotted, and dashed lines show predicted WD cooling
sequences for CO core and H envelopes with stellar masses equal to 0.5, 0.7, and 0.9 M/M⊙ (Althaus, & Benvenuto 1998). Theoretical
predictionshavebeentransformedintotheobservationalplanebyadoptingHatmospheremodels. Diamondsshowa12Gyrisochronefor
Z=0.001 and Y=0.232constructed by Cariuloet al. (2004, M ≥0.5M⊙)and by Baraffe et al. (1997, M <0.5M⊙). Panel b) -Same as
panela),butforWDcoolingsequences,withCOcoreandHeenvelopes(Benvenuto, &Althaus1997),transformedintotheobservational
plane by adopting He atmosphere models. Panel c) - Same as panel a), but for WD cooling sequences with He core and H envelopes
(Serenellietal. 2002), stellarmassesequalto0.30,0.39, and0.45M/M⊙ andHatmospheres.
8 Monelli et al.
Fig. 4.— Color-Magnitudediagram inthe R,Hα−R bands. Predicted cooling sequences are the sameas inFig. 4. Thethick dashed
linemarksthedetection limit. Errorbarsdisplayintrinsicphotometricerrors.