Near Infrared Survey of Populous Clusters in the LMC: Preliminary Results

Abstract

We report preliminary results from our near-infrared JHK survey of star clusters in the LMC. The primary goals of the survey are to study the three-dimensional structure and distance of the LMC. In 2003 and 2004 we used the Infrared Side Port Imager (ISPI) on the CTIO 4m to obtain near infrared photometry for a sample of populous LMC clusters. We utilize the K-band luminosity of core helium burning red clump (RC) stars to obtain individual cluster distances and present a preliminary assessment of the structure and geometry of the LMC based on a subset of our data.

Full text

Page 1

arXiv:astro-ph/0506760v1 30 Jun 2005
Resolved Stellar Populations
ASP Conference Series, Vol. TBA, 2005
D. Valls–Gabaud & M. Chavez (eds)
Near Infrared Survey of Populous Clusters in the LMC:
Preliminary Results
A.J. Grocholski1,2, A. Sarajedini1, K.A.G. Olsen3, and G.P. Tiede4
1Department of Astronomy, University of Florida, 211 Bryant Space
Science Center, Gainesville, Florida 32611, aaron@astro.ufl.edu,
ata@astro.ufl.edu
2Visiting Astronomer, Cerro Tololo Inter-American Observatory
3Cerro Tolo Inter-American Observatory, National Optical Astronomy
Observatory, Casilla 603, La Serena, Chile, kolsen@ctio.noao.edu
4Department of Physics and Astronomy, Bowling Green State
University, Bowling Green, OH 43403, gptiede@bgnet.bgsu.edu
Abstract.
We report preliminary results from our near-infrared JHK survey of star
clusters in the LMC. The primary goals of the survey are to study the three-
dimensional structure and distance of the LMC. In 2003 and 2004 we used the
Infrared Side Port Imager (ISPI) on the CTIO 4m to obtain near infrared pho-
tometry for a sample of populous LMC clusters. We utilize the K-band lumi-
nosity of core helium burning red clump (RC) stars to obtain individual cluster
distances and present a preliminary assessment of the structure and geometry
of the LMC based on a subset of our data.
1.Introduction
The Large Magellanic Cloud (LMC) is an attractive astronomical target for
a variety of reasons. Owing to its relative proximity, stellar populations in the
LMC can be easily resolved. These populations exhibit an array of star formation
processes and episodes in a dynamic environment making the LMC well suited
for studying the formation and evolution of a satellite galaxy. Traditionally, the
LMC has been thought of as an approximately planar galaxy that, in spite of
its proximity, can be assumed to lie at a single distance from us. In contrast,
Caldwell & Coulson (1986) have shown that the LMC disk is tilted with respect
to the plane of the sky. More recent work has not only confirmed that the LMC
is tilted, but it also indicates that the LMC disk is considerably thicker than
previously assumed.
van der Marel & Cioni (2001) determined, through the use of red giant
branch and asymptotic giant branch stars as relative distance indicators, that
the LMC is tilted 34.◦7±6.◦2 with respect to the line of sight (0◦is face-on) such
that the Northeast portion of the LMC is closer to us than the Southwest. Olsen
& Salyk (2002) confirmed this result, finding i = 35.◦8±2.◦4 by utilizing core He
burning red clump (RC) stars as their relative distance indicator. Additionally,
1

Page 2

2
Grocholski et al.
by studying carbon star kinematics in the LMC disk, van der Marel et al. (2002)
have determined that v/σ = 2.9 ± 0.9, implying that the LMC disk is thicker
than the Milky Way thick disk (v/σ ≈ 3.9). Lastly, the Magellanic Stream
(e.g. Putman et al. 2003), flaring (Alves & Nelson, 2000) and elongation of the
LMC disk (van der Marel & Cioni 2001) and the possibility that the LMC disk
is warped (Olsen & Salyk 2002, Nikolaev et al. 2004) all indicate that the LMC
has not escaped unharmed from its tidal interactions with the Milky Way and
Small Magellanic Cloud.
The distance to the LMC has been a topic of considerable discussion in
recent years and a variety of methods have been employed to calculate this
distance; e.g. variable stars (Cepheids, RR Lyraes, Miras), color-magnitude di-
agram (CMD) features (main sequence turn off, tip of the red giant branch, RC
stars), and SN 1987a. There has been, until recently, little agreement between
the different methods and sometimes even amongst distances calculated using a
single method. This lead to a “long” and “short” distance scale for the LMC
with a “short” distance modulus, (m − M)0, of ∼18.2-18.3 mag and a “long”
distance of ∼18.5-18.7 mag. Clementini et al. (2003) demonstrate this distance
problem (top panel, their Fig. 8), and find that the long and short distance
scale can be reconciled, at least to within the errors, with improved photometry
and/or reddening estimates (bottom panel, their Fig. 8) for some of the previous
works.
A primary reason for interest in the LMC distance is its use as the extra-
galactic distance scale zeropoint. The Hubble Space Telescope Key Project to de-
termine H0(Freedman et al. 2001) utilized a sample of LMC Cepheid variables to
define the fiducial period-luminosity relation. Cepheid distances were then used
to calibrate secondary standard candles which lie further along the extragalactic
distance ladder. Thus, the accuracy of their determination of H0(72±8 km s−1
Mpc−1) hinges on the accuracy of their zeropoint, (m−M)0,LMC= 18.5±0.10.
It turns out the error in their calculation is dominated by the uncertainty in
(m − M)0,LMC; it takes up 6.5% of their 9% error budget (Mould et al. 2000).
In this paper we will present preliminary results from our near-infrared
survey of populous clusters in the LMC. Section 2 presents our observations of
LMC cluster and field stars. In the next two sections we discuss our application
of the K-band luminosity of the RC as a standard candle for calculating absolute
cluster distances (§3) and for determining relative distances to the LMC fields
(§4). Finally, in §5 we talk about our future work on this project.
2.Data
2.1.Observations
We have obtained near infrared images for a sample of intermediate age LMC
clusters over the course of two, three night observing runs (20-22 January 2003
and 06-08 February 2004) at the CTIO 4m. Our observations were made with
the Infrared Side Port Imager (ISPI) which utilizes a 2048 × 2048 pixel HAWAII
2 HgCdTe array. In the f/8 configuration, ISPI yields an 11′× 11′field of view
and a plate scale of ∼ 0.′′33 pixel−1. For our observations we used a nine-point
dither pattern, centered on each cluster, and total integration times as follows: J
= 540s, H = 846s, K′= 846s. Average seeing for all six nights was ∼ 1.2′′. Table

Page 3

Populous Clusters in the LMC
3
1 lists two of our 18 program clusters along with right ascention and declination
(J2000) and passbands in which the clusters were observed.
Table 1. LMC Cluster Sample
Cluster Alternate nameRA (J2000)Dec (J2000)Filters
NGC 1651
Hodge 4
04h37m32.s67
05 32 25.00
-70◦35′07.′′7
-64 44 12.0
JHK′
JHK′
SL 556
2.2.Reduction
All data was processed using standard data reduction steps, which we will now
summarize. Images were dark subtracted, sky subtracted and then flat fielded
using on-off dome flats. Due to the combination of ISPI’s wide field of view and
the relatively large steps in our dither pattern (≥ 30′), these images suffer from
geometric distortions, caused mostly by the curvature of the focal plane. To
correct for this, we apply a high order distortion correction to each image using
the IRAF task GEOTRAN. The corrected images are then aligned, shifted and
averaged to create a final science frame for each cluster and filter.
Science frames were photometered using a combination of DAOPHOT and
ALLSTAR (Stetson 1987) as follows. A rough PSF was constructed using the
brightest ∼ 200 stars in each image. The rough PSF was then used to remove
neighbors from around the PSF stars, allowing the creation of a more robust
PSF from the cleaned image. ALLSTAR was utilized to fit this improved PSF
to all stars in the science frame. In an effort to detect and photometer faint stars
and/or companions, we used a single iteration of subtracting all stars detected
and fit in the first ALLSTAR pass, then searching for previously undetected stars
in our fields. All new detections were run through ALLSTAR using the same
PSF as in the first ALLSTAR pass and these stars were added to the photometry
list. At this point, aperture corrections, calculated for each frame, were applied
to the PSF photometry. Finally, aperture corrected photometry lists from each
filter were combined with the requirement that a star be detected in all filters for
it to be kept in the final combined list of instrumental magnitudes. Zero points
and color tranformations appropriate for our data were calculated by comparing
our instrumental magnitudes with photometry from the 2MASS All-Sky Data
Release1for each field in our program.
3.Preliminary Distances
Figure 1 presents (K, J −K) CMDs for NGC 1651 and Hodge 4, where all stars
within ∼ 1′of the cluster centers are shown. Both CMDs show a prominent RC
at K ∼ 16.9 and a well populated RGB extending up to K ∼ 12.5.
We follow the method of Grocholski & Sarajedini (2002) in using a box that
extends 0.8 mag in K and 0.2 mag in J −K (shown in Fig. 1), centered by eye,
1http://www.ipac.caltech.edu/2mass/releases/allsky

Page 4

4
Grocholski et al.
Figure 1.
show a well populated helium burning RC, as denoted by the box, along with
an RGB extending ∼ 4.5 mag brightward of the RC. Median values of KRC
are given.
Near infrared CMDs for NGC 1651 and Hodge 4. Both plots
to select the RC stars. KRCis calculated by taking the median value of all stars
within this box. For NGC 1651, KRC= 16.93 ± 0.02 and KRC= 16.81 ± 0.02
for Hodge 4.
With regards to the reddening of each cluster, we utilize the dust maps of
Burstein & Heiles (1982) and Schlegel, Finkbeiner, & Davis (1998). Since the
values determined from both dust maps, for each cluster, are in good agreement,
we adopt the average value from the two maps as our cluster reddenings. For
NGC 1651 we find E(B − V ) = 0.12 ± 0.02 and for Hodge 4, E(B − V ) =
0.05±0.01. Using the relations from Cardelli, Clayton, & Mathis (1989), AV =
3.1E(B−V ) and AK= 0.11AV, these reddenings translate to AK= 0.041±0.003
and AK= 0.017 ± 0.007.
Previous authors have shown that the absolute RC magnitude varies as
a function of age and metallicity for visible and near-infrared bands, with this
variation seen in both theoretical (e.g. Salaris & Girardi 2002) and observational
data (e.g. Sarajedini 1999, Cole 1998). An in-depth comparison of the absolute
K-band RC magnitude with age and metallicity for a sample of simple stellar
populations was performed by Grocholski & Sarajedini (2002). These authors
advocate using an interpolation over either their observational data or the theo-
retical models of Girardi & Salaris (2001) to create an MK(RC) “plane” which,
given a cluster’s age and metallicity, can be used to predict MK(RC) for that
cluster.
In many cases, however, age and metallicity values for LMC clusters are
not readily available or not reliable. For example, Olszewski et al. (1991) have
presented the only large scale determination of metallicities for LMC clusters,
based on the Ca II triplet. However, many of their cluster [Fe/H] values, in-

Page 5

Populous Clusters in the LMC
5
cluding NGC 1651 and Hodge 4, are based on observations of a single star.
Sarajedini et al. (2002) calculated [Fe/H] values for both of these clusters us-
ing the slope of the RGB. While their value for Hodge 4 is consistent with
that of Olszewski et al. (1991), their value for NGC 1651 is 0.3 dex more
metal rich.As such, in the current work we choose not to apply the full
calibration of MK(RC) as discussed above, but rather we adopt the value of
MK(RC) = −1.61 ± 0.04 given in Grocholski & Sarajedini (2002). We note
that Sarajedini et al. (2002), utilizing the full RC treatment from Grochol-
ski & Sarajedini (2002), find MK(RC) = −1.56 ± 0.12 for NGC 1651 and
MK(RC) = −1.64 ± 0.17. This implies that error in the value we have cho-
sen to use for MK(RC) is likely larger than that quoted.
Using the values listed above for K(RC), MK(RC), and AKfor each custer,
we find for NGC 1651, (m−M)0= 18.50±0.06 and (m−M)0= 18.40±0.05 for
Hodge 4. The errors quoted are the random errors added in quadrature. These
numbers are consistent with the LMC distance, (m − M)0= 18.50 ± 0.10 used
in the HST Key Project to determine an accurate value of H0(see Freedman
et al. 2001 for more information). Additionally, these distances agree with the
LMC geometry determined by van der Marel & Cioni (2001) and Olsen & Salyk
(2002) in that Hodge 4 should be closer to us than NGC 1651, based on the tilt
of the LMC’s disk and the location of the clusters in the LMC. Lastly, Sarajedini
et al. (2002) find, for NGC 1651 and Hodge 4, (m − M)0= 18.55 ± 0.12 and
18.52 ± 0.17. These distances are in agreement, within the errors, with our
results.
4. Field Stars
In Figure 2 we show (K, J−K) CMDs for the field stars surrounding our clusters
and, as with the clusters, the RC and RGB for the two fields are easily visible.
The major difference between the cluster and field CMDs is the wider field
RGB (spread in color) and larger field RC (spread in both color and luminosity)
caused by the intrinsic distribution in age and metallicity of the field population.
Although this situation is a bit more complicated for the RC (see Salaris &
Girardi 2002; Grocholski & Sarajedini 2002), in general, older and/or more
metal rich populations have redder RGBs and redder and fainter RCs than
young and/or metal poor populations. Due to the mixed population in the
LMC field, dealing with the field RC luminosity as a standard candle becomes
a much more formidable task than dealing with a simple stellar population RC.
However, if we can make the assumption that each observed field within the
LMC has a similar mix of stars in terms of age and metallicity, then the field RC
can be easily applied as a relative distance indicator since MK(RC) should be
the same for all fields. This assumption is reasonable given that the bulk of the
LMC RC stars are ∼ 4 Gyr old (Girardi & Salaris 2001) and differential rotation
should destroy any record of the initial local age and metallicity distribution on
timescales much shorter than this (Olsen and Salyk 2002). Additionally, if this
assumption holds true, the variation in RC colors amongst our fields is indicative
of the relative reddenings of these fields.
Similar to §3, we determine the magnitude and color values of the field RC
by taking the median value of all stars within a predefined box, centered on the