Spectroscopic study of star clusters in the Small Magellanic Cloud

Spectroscopic study of star clusters in the Small Magellanic Cloud

Asociación Argentina de Astronomı́a
BAAA, Vol. 48, 2005
E.M. Arnal, A. Brunini, J.J. Clariá Olmedo, J.C. Forte, D.O. Gómez, D. Garcı́a Lambas,
Z. López Garcı́a, S. M. Malaroda, & G.E. Romero, eds.
COMUNICACIÓN DE TRABAJO – CONTRIBUTED PAPER
Spectroscopic study of star clusters in the Small
Magellanic Cloud: star formation history
Juan J. Clariá
Observatorio Astronómico, Universidad Nacional de Córdoba,
[email protected]
Andrés E. Piatti
IAFE, Buenos Aires, Argentina, [email protected]
João F.C. Santos Jr.
Dpto. de Fı́sica, UFMG, Belo Horizonte, Brasil, [email protected]
Eduardo Bica
Instituto de Fı́sica, UFRGS, Porto Alegre, Brasil, [email protected]
Andrea V. Ahumada and M. Celeste Parisi
Observatorio Astronómico, Universidad Nacional de Córdoba, andrea,
[email protected]
Abstract. We present integrated spectra in the range (3600-6800) Å for
18 concentrated clusters of the Small Magellanic Cloud (SMC). By using
the template matching and equivalent width methods, ages and metallicities were respectively determined and a very good agreement between
ages derived from both methods was found. Combining the present sample with 19 additional SMC clusters whose ages and metallicities were put
onto a homogeneous scale, we analyze the clusters’ age and metallicity
distributions. By considering the deprojected distances of the clusters
from the SMC center instead of their projections onto the right ascension
and declination axes, the ensuing analysis seems to indicate that the SMC
inner disk could have been the result of a cluster formation episode which
reached a peak ∼ 2.5 Gyr ago. Evidence for a metallicity gradient in the
SMC disk is also presented.
Resumen. Presentamos espectros integrados en el rango (3600-6800)
Å de 18 cúmulos concentrados de la Nube Menor de Magallanes (NMM).
Las edades y metalicidades fueron derivadas a partir del método de ajuste
de templates y de los anchos equivalentes, respectivamente. Las edades
obtenidas por ambos métodos muestran muy buen acuerdo. Combinando
la presente muestra con 19 cúmulos adicionales para los cuales las edades y
metalicidades fueron determinadas en una escala homogénea, analizamos
140
Star clusters in the SMC: star formation history
141
las distribuciones de edad y metalicidad de los cúmulos. Considerando las
distancias deproyectadas de los cúmulos al centro de la NMM, en lugar
de sus proyecciones sobre los ejes de ascensión recta y declinación, este
análisis sugiere que el disco interior de la NMM podrı́a haberse formado
como consecuencia de un episodio de formación estelar ocurrido hace ∼
2.5x109 años. Presentamos además evidencia sobre la existencia de un
gradiente de metalicidad en el disco de la NMM.
1.
Introduction
In order to analize star clusters in dwarf galaxies which can be observed through
ground-based large telescopes as well as through the Hubble Space Telescope,
a star cluster spectral library at the SMC metallicity level can prove useful.
We present here integrated spectra for 18 concentrated SMC clusters. Previous
cluster samples are complemented by the present one in an attempt to provide a
spectral library with several clusters per age bin. At the same time, the cluster
parameters are determined and the age and metallicity distributions analized.
This aims at looking into the SMC star formation history and the chemical
enrichment processes.
2.
Spectroscopic observations
By using the CASLEO (Argentina) 2.15 m and CTIO (Chile) 1.5 m telescopes,
integrated spectra for a sample of 18 concentrated SMC clusters were obtained
during 13 observing nights. The spectral coverage was (3600-6800) Å, while the
typical resolution and dispersion were 12 Å and 3.5 Å/pixel, respectively. The
observations were performed by scanning the slit across the objects in the NorthSouth direction.
3.
The template-matching and equivalent widths methods
We resorted to two independent methods to derive the cluster parameters: the
template-matching method, which entails comparing and matching the observed
spectra to template spectra with well-determined properties (e.g., Piatti et al.
2002a), and the equivalent width (EW) method, which consists in utilizing diagnostic diagrams involving the sum of EWs of selected spectral lines, along
with their calibrations with age and metallicity (Santos & Piatti 2004, hereafter
SP). Cluster reddening values were first estimated by interpolation between the
extinction maps of Burstein & Heiles (1982) and Schlegel et al. (1998). Before measuring EWs, the spectra were set to the rest-frame according to the
Doppler shift of H Balmer lines. The EWs of H Balmer, K CaII, G band (CH)
and MgI (5167+5173+5184) Å were measured using IRAF task splot. Subsequently, the sum of EWs of the 3 metallic lines (S m ) and of the 3 Balmer lines
Hβ, Hγ and Hδ (Sh ) were used to estimate cluster parameters. Summing up,
the calibrations, aided by diagnostic diagrams involving S m and Sh , allowed us
to derive age for star clusters younger than ∼ 10 Gyr and metallicity for older
ones. For intermediate-age clusters (2.5 < t(Gyr) < 10) with [Fe/H] > -1.4, it
142
J.J. Clariá et al.
is necessary to constrain age by using the template-matching method and then
obtain metallicity with the SP’s calibration, if the cluster is old. We found a
very good agreement between ages derived from both methods. As examples, we
show in Fig. 1 the best template combination for K5, i.e., the average of Ia and
Yh templates (equivalent to 0.8 Gyr), compared with the reddening-corrected
spectrum, and K3, i.e., the average of G3 and Ia templates (equivalent to 7
Gyr), compared with the reddening-corrected spectrum. For K5, its metallicity
has been corrected to [Fe/H] = -0.5, following an age revision on the Piatti et
al. (2005) value. Brocato et al. (2001) presented a HST CMD of K3 making
its photometry available, on which we have superimposed Padova isochrones
(Girardi et al. 2002) to obtain [Fe/H] = -1.2 and t = 6 Gyr. Good agreement was reached between age and metallicity values obtained in the present
analysis for K3 and the results of previous studies (see, e.g., Mighell et al. 1998).
The derived ages and metallicities for the cluster sample are summarized in Table 1. The methods used to obtain age and metallicity are indicated in columns
6 and 9. All clusters were age-ranked according to the EW method with the
only exception of K28, with a low S/N spectrum. The template method was
applied to the whole sample either independently from the EW method (minus
sign in column 6) or in conjunction with the EW method (plus sign in column 6).
The last two columns of Table 1 show the cluster metallicities adopted whenever
possible as well as their corresponding sources, respectively. For K3, we used
equation (8) of SP. We also fitted Padova isochrones to the K6 CMD obtained
by Matteuci et al. (2002) and yielded a cluster metallicity of [Fe/H] = -0.7, assuming the SMC apparent distance modulus (m-M) = 19.0 (Cioni et al. 2000).
Table 1.
Cluster
Cluster parameters
tliterature
Ref.
tm
method
tadopted
[Fe/H]
Ref.
(Gyr)
(Gyr)
(Gyr)
L5
0.03
4.1
1
0.8
Sh ,Sm - template
3.0±1.5
-1.1±0.2
1,9
K5
0.02
2.0
1
0.8
Sh ,Sm - template
1.2±0.5
-0.5±0.2
1,9
K3
0.02
7.0±1.0
2,4
7.0
Sh ,Sm + template
7.0±1.0
-1.20±0.2
2,9
K6
0.03
1.3
7
2.0
Sh ,Sm + template
1.6±0.4
-0.7
9
K7
0.02
3.5
3
4.0
Sh ,Sm + template
3.5±0.5
-1.0
3
HW 8
0.03
0.05
Sh ,Sm - template
0.05±0.02
NGC 269
0.01
0.6
Sm - template
0.6±0.2
L 39
0.01
0.015
Sh ,Sm - template
0.015±0.010
K 28
0.06
2.1
2
1.0
template
1.5±0.6
-1.0±0.2
2,9
NGC 294
0.02
0.3
Sh ,Sm - template
0.3±0.1
L 51
0.07
0.015
Sh ,Sm - template
0.015±0.010
K 42
0.06
0.045
Sh ,Sm - template
0.045±0.015
L 66
0.06
0.015
Sh ,Sm - template
0.015±0.010
NGC 411
0.03
1.5±0.2
2,6,8
1.0
Sh ,Sm - template
1.5±0.3
-0.7±0.2
2,6,8
NGC 419
0.03
1.6±0.4
2,4
0.8
Sh ,Sm - template
1.2±0.4
-0.7
2
NGC 422
0.03
0.3
Sh ,Sm - template
0.3±0.1
IC 1641
0.03
0.3
Sh ,Sm - template
0.3±0.1
NGC 458
0.02
0.17±0.03
2,5
0.05
Sh ,Sm - template
0.13±0.06
-0.23
2
References: (1) Piatti et al. (2005); (2) Piatti et al. (2002b); (3) Mould et al. (1992); (4) Rich et al. (2000); (5)
Alcaino et al. (2003); (6) Alves & Sarajedini (1999); (7) Matteucci et al. (2002); (8) Leonardi & Rose (2003); (9) this
work.
4.
E(B − V )
Cluster age and metallicity distributions
In order to examine the cluster age and metallicity distributions, we computed
for each cluster its deprojected distance from the SMC optical center using the
expression:
Star clusters in the SMC: star formation history
143
d = d(p)[1 + [sen(p − p0 )2 ][tang(i)]2 ]0.5 , (1)
where d is the deprojected distance from the SMC optical centre, d(p) the projected distance on the plane of the sky, p the position angle of the cluster, p 0 the
position angle of the line of nodes and i the tilt of the SMC plane to the plane
of the sky. In equation (1), the factor sen(p − p 0 )2 considers how distant the
object lies from the line of nodes. If the object were lying over the line of nodes,
d and d(p) will coincide, regardless of the disk tilt. The SMC optical centre was
assumed to be placed (J2000) at: α = 00 h 52m 45s , δ = -72◦ 49’ 43” (Crowl et
al. 2001). To compute d from equation (1), we used p 0 = 45◦ and i = 60◦ (de
Vaucouleurs & Freeman 1973).
The left panel of Fig. 2 shows how the derived ages vary as a function of the
computed deprojected distances. There are very few clusters younger than 4
Gyr in the outer disk, defined as the portion of the SMC disk with d ≥ 3.5 ◦ ,
while there are very few clusters older than 4 Gyr in the inner disk. What is
more, in the inner disk, as the clusters become older, their corresponding deprojected distances get proportionally larger, a fact which surprisingly suggests the
possibility that the clusters may have been formed outside in, as in a relatively
rapid collapse.
Piatti et al. (2005) confirmed that ∼ 2.5 Gyr ago the SMC reached the peak of a
burst of cluster formation, which corresponds to a very close encounter with the
LMC, according to the recent dynamical models of Bekki et al. (2004). Piatti et
al. (2005) studied 10 clusters with ages and metallicities in the ranges 1.5-4 Gyr
and -1.3 < [Fe/H] < -0.6, respectively. They were more inclined to believe in
a bursting cluster formation history rather than a continuous one for the SMC.
The age-position relation shown in Fig. 2 (left-hand panel) for clusters younger
than 4 Gyr adds, if confirmed, a new nuance to the bursting theory of cluster
formation. In the cluster formation episode peaking at ∼ 2.5 Gyr, the burst
could have originated the formation process which continued producing clusters
from the outermost regions to the innermost ones in the inner SMC disk. Under
these circumstances, the inner disk could have been formed during this period.
Fig. 2 (right panel) depicts the distribution of the cluster metallicities as a
function of the deprojected distances from the SMC center. Note that in the
outer disk, there are no clusters with [Fe/H] values larger than ∼ -1.2, with
only one exception. The inner disk, however, is shared by both metal-poor and
metal-rich clusters, the averaged metallicity being clearly larger than that for
the outer disk. We then confirm the existence of a metal abundance gradient for
the SMC disk, given the fact that the farther a cluster from the galaxy centre,
the poorer its metal content. Nonetheless, all the clusters with [Fe/H] > -1.2
in the inner disk were formed during the last 4 Gyr, whereas the metal-poor
ones are as old as those in the outer disk. Accordingly, the abundance gradient
seems to be representative of the combination of an older and more metal-poor
144
J.J. Clariá et al.
Figure 1.
Integrated spectra of K5 (left-hand panel) and K3 (righthand panel), corrected for reddening, and the templates which best
match them.
population of clusters spread all through the SMC and a younger and metalricher one predominantly born in the inner disk. Note that some few clusters
were also formed in the inner disk with [Fe/H] ∼ -1.2.
References
Alcaino G., Alvarado F., Borissova J. & Kurtev R. 2003, A&A, 400, 917
Alves D.R. & Sarajedini A. 1999, ApJ, 511, 225
Bekki K., Couch W.J. et al. 2004, ApJ, 610, L93
Burstein D. & Heiles C. 1982, AJ, 87, 1165
Brocato E., Di Carlo E. & Menna G. 2001, A&A, 374, 523
Cioni M.R., van der Marel R.P., Loup C. & Habing J. 2000, A&A, 359, 601
Crowl H.H., Sarajedini A., Piatti A.E. et al. 2001, AJ, 122, 220
de Vaucouleurs G. & Freeman K.C. 1973, Vistas in Astron., 14, 163
Girardi L., Bertelli G. et al. 2002, A&A, 391, 195
Leonardi A.J. & Rose J.A. 2003, AJ, 126, 1811
Matteucci A., Ripepi V., Brocato E., Castelani V. 2002, A&A, 387, 861
Mighell K.J., Sarajedini, A. & French R.S. 1998, AJ, 116, 2395
Star clusters in the SMC: star formation history
Figure 2.
Left: Cluster ages versus deprojected distances. Studied
clusters (crossed boxes) and 19 additional clusters (triangles) taken
from Piatti et al. (2002b) and Piatti et al. (2005) are indicated. Right:
Cluster metallicities versus deprojected distances.
Mould J.R., Jensen J.B. & Da Costa G.S. 1992, ApJS, 82, 489
Piatti A.E., Bica E., Clariá J.J. et al. 2002a, MNRAS, 335, 233
Piatti A.E., Sarajedini A., Geisler D. et al. 2002b, MNRAS, 329, 556
Piatti A.E., Sarajedini A., Geisler D. et al. 2005, MNRAS, 358, 1215
Rich R.M., Shara M., Fall S.M. & Zurek D. 2000, AJ, 119, 197
Santos Jr., J.F.C. & Piatti A.E. 2004, A&A 428, 79 (SP)
Schlegel D., Finkbeiner D. & Davis M. 1998, ApJ, 500, 525
145