VO Study of Super Star Clusters in Nearby Galaxies

1
VO Study of Super Star Clusters in Nearby Galaxies

• Brad Whitmore
• IAU Meeting, August 18, 2006

• OUTLINE
• Science Background and Goals
• VO Connection
• Interpretation The Big Picture

2
Science Background and Goals

• The discovery of young compact star clusters
  in merging galaxies, such as the Antennae
  Galaxies , has revitalized the study of star
  clusters.
• In particular, the brightest clusters have
  all the attributes expected of a young globular
  cluster.
• Hence, it is now possible to study the
  formation of globular clusters in the local
  universe rather than trying to figure out how
  they formed some 13 billion years ago.

3
(No Transcript)
4
(No Transcript)
5

• Roughly 20 gas-rich mergers have now been
  observed in detail by HST. All show young star
  clusters.
• In addition, we find young, massive, compact
  clusters in
• starburst dwarf galaxies (e.g.,
  Meurer et al., 1995),
• barred galaxies (Barth et al.,
  1995),
• spiral galaxies (Larsen Richtler,
  1999)
• Milky Way and LMC (e.g., Walborn
  2000)
• These clusters have properties similar to
  those seen in the mergers, but always fewer in
  number, and generally fainter in luminosity.
• Primary Science Question Is violent
  star cluster formation (e.g., in mergers)
  different than quiescent star formation
  (e.g., in spirals)?

6
(No Transcript)
7
Whitmore, 2000,2003
If there are two different modes of star cluster
formation, we might expect a bimodal distribution
in a plot of the magnitude of the brightest
cluster in a galaxy vs. the log of the number of
clusters.
Violent star formation ?
Quiescent star formation ?
8
Whitmore, 2000,2003

• The data appear to support a universal model
  rather than a bimodal model, with the correlation
  being due to statistics, not physics.
• However, this dataset, and the reductions, were
  very inhomogeous.
• Our goal is to redo this diagram
• - with a uniform data set (e.g., SDSS, HST)
• with uniform analysis (e.g., WESIX, Sextractor)
• - for larger dataset (e.g., N 100)

Best fit
Predicted if universal power-law, index -2
9
VO Connection

• This project grew out of a student project
  developed at the 2005
• NVO Workshop in Aspen, Colorado.
• Following the workshop we submitted a proposal
  for a NVO
• Grant, which was successful.
• Co-Investigators are K. Lindsay and C. Hanley
  (STScI), B. Chan (IPAC) and R. Chandar (JHU,
  Carnegie Obs.)

10
Feasibility Study M51 (using WESIX)
DataScope - SDSS g-band image from WESIX - Source
extraction and cross matching ALADIN
visualization Voplot analysis
11
Photometric Calibration
Compared SDSS g-mag from sextractor to HST V-mag
(Rupali Chandar) Scatter 0.1mag
12
Analysis with VOplot
Source classification with flux concentration
index (aperture mag isophot mag) VOTables
exported back to Aladin for various source types
13
Diffuse sources
Nucleus
Clusters
Saturated stars
Compact objects (stars)
14
Compact objects (stars)
15
Nucleus
Saturated stars
16
Diffuse sources
17
Clusters
18
Fraction of missing clusters Red crosses 2
mass Blue squares clusters Fraction hidden by
dust (outside center) lt 45 (15/33) NOTE -
Something different near center !
19
Whitmore, 2000,2003
The single datapoint we established at the Aspen
workshop appears to agree with the universal
interpretation.
Best fit
M51
Predicted if universal power-law, index -2
20

• One change we made since the Aspen prototype
  study was to use sextractor directly in order to
  linearize the clasification diagram (using the
  difference between two aperture sizes a
  capability not provided by WESIX).
• Other things we have learned.
• - need to adjust selection criteria to remove
  stars for each target.
• can only reach to (m-M) 30 (10 kpc) with SDSS.
• - will need to look at using HST preview images
  for more distant galaxies.

21
Here is a version of the plot including our new
data (in blue). There appears to be a systematic
downward shift. This may indicate that the
earlier studies were very incomplete at faint
magnitudes.
Best fit
M100
M51
M101
4258L
3194
M66
4214
4258R
3351
3521
5585
M63
M94
M81
3368
3623
5204
Predicted if universal power-law, index -2
22
The Big Picture - A General Framework for
Understanding the Demographics of Star Clusters

• Ingredients (assume all stars form in
  clusters)
• A universal initial mass function (power law,
  index -2)
• 2. Various star(cluster) formation histories
• 3. Various cluster disruption mechanisms
• (e.g., T-1 lt 100 Myr, 2-body relaxation gt
  100 Myr)
• 4. Convolution with observational artifacts and
  selection effects
• Observations (luminosity and age distributions,
  color-color diagrams, total luminosity of a
  galaxy, fraction of field stars, )

23
Here is an example of two of our models based on
observations of the Antennae. Whitmore,
Chandar, Fall (2006)
24
We have used this model to simulate the
Mv(brightest) vs. Log N diagram for different
values of the power law index (using the old
dataset for now). We find that values in the
range 2.0 to -2.4 fit the data well. However,
the new data may be slightly shallower, hence
there may be a second order difference between
spirals and mergers. Whitmore, Chandar Fall
2006
25
The model is also able to explain the scatter in
the diagram (confirming a previous finding by
Larsen 2004). Whitmore, Chandar Fall 2006
26
Summary

• VO Tools (e.g., datascope, WESIX, Aladin, VOPLOT)
  were very useful for prototyping the reductions
  for our study. We eventually used sextractor
  directly to have more flexibility.
• SDSS images can be used for this project
  (although we plan to use HST preview images to
  expand the sample).
• We have developed a Monte-Carlo simulation to
  better interpret the Mv(brightest) vs. Log N
  diagram.