6 The cluster population

As stated in Sect. 2, we have two samples of star clusters: those detected in all three WFPC2 filters, and those only detected in the F555W and F814W filters. We will discuss each set separately, beginning with the three-filter set. For this set of objects we have greater age discrimination, and we will use this information to investigate cluster age spreads both within and between the clumps.

6.1 F255W, F555W, and F814W

Figure 7: A color map made from the F555W and F814W images shows bluer objects as darker, and redder as lighter. Orientation is as for Fig. 4. Clump B is prominent in the top left (NE) corner of this image; Clumps C and F are also very blue. This image also displays the complex small-scale dust structures associated with the nuclear region (Clump A).

Figure 8: This F555W filter image of the central part of NGC 7673 shows the pair of linear dust structures emanating from a bright source, which may be the nucleus. Orientation is as for Fig. 4.

Figure 10: Two color plots for the clusters. Filled stars mark ages of 1, 5, 10, 20, 50, 100, and 500 Myrs on a model track for purely stellar emmission (see Sect. 4.1). Clockwise from top left: nuclear, the "blue blob'' which is Clump B, all other clusters not associated with a particular clump, and Clump C. The arrow indicates Av=1. See discussion in the text.

We have detected 50 objects with our selection criteria in the three filters: 12 in the nuclear region, 8 in Clump B, 12 in Clump C, and 18 others not associated with a particular clump. The coordinates, colors, and F555W magnitudes are presented in Table 2, and Fig. 9 shows the objects overplotted on a F555W image as white dots with number labels. Two-color diagrams for these objects are shown separated by region in Fig. 10, with an overplotted Starburst99 cluster model track for Z=0.008, with ages in Myr. The reddening line is for A_v=1.0 using a Calzetti et al. (1994) extinction law for starburst galaxies. Mean [255-555] and [555-814] colors are presented in Table 3.

From the color map in Fig. 7, we would not expect all of the clusters to escape the effects of dust extinction. However, we might expect to see only weak dust effects in our sample, because we are heavily biased in favor of clusters which do not have significant reddening (for a discussion see Conselice et al. 2000b and Calzetti 2001). One magnitude of visual extinction makes a cluster approximately 2.4 mag fainter in our F255W filter. Thus we only expect to see the brightest, so the youngest and most massive, star clusters in this sample. At its most luminous, a $10^{6}~M_{\odot}$cluster is predicted to have a F255W magnitude of -17.9 at an age of 3 Myr, or an apparent m(F255W) $\approx 15.3$ at the distance of NGC 7673. With our selection criteria we are sensitive to 3 Myr clusters down to $10^{4}~M_{\odot}$, but only $10^{5}~M_{\odot}$ clusters at 20 Myr, due to rapid fading with age.

6.1.1 Nuclear region

The nuclear region is unique as the only area significantly affected by dust in this galaxy. This complicates our interpretation of the colors, because of the degeneracy between age and reddening. The reddening line can place a cluster at 1 Myr on the track anywhere from 2 to 100 Myr.

The nuclear clusters span a wide range of [255-555] and [555-814] colors. Three clusters (Numbers 4, 5, and 7) have blue [555-814] and [255-555] colors, and we suggest that these are young, $\sim$6 Myr clusters suffering from little extinction. Four clusters (Numbers 1, 8, 11, and 12) concentrate near [255-555] $\sim$ -1.1 and [555-814] $\sim$ 0.45. This places them on the model track at 30-50 Myr. A_v=0.5 could make 20 Myr clusters appear here, and A_v=0.8 could cloak 5 Myr clusters and move them to the 30-50 Myr position on the two-color diagram. Three of the clusters (Numbers 2, 9, and 10) have [255-555] less than -0.2 and [555-814] $\sim$ 0.6. These could be 200 Myr with a small amount of reddening, or 7-50 Myr with A_v=1.0.

Figure 11: Magnitude-color plot for the 3-filter selected clusters shown in Figs. 9 and 10. The 12 nuclear clusters are represented by triangles, the 8 Clump B clusters by squares, the 12 Clump C clusters with diamonds, and the others not associated with a particular region with stars. The model track is for an instantaneous burst of $10^{6}~M_{\odot}$ with Z= 0.008. With an assumption of age, we can shift the model track vertically and derive a mass for the cluster based on its F555W luminosity. The age assumption is most secure for the Clump B clusters (<7 Myr); thus the brightest clusters in Clump B have masses around $5 \times 10^{5}~M_{\odot}$.

 

 

Table 3: Mean [255-555] and [555-814] cluster colors for the three-filter selected sample.

Region	Number of Members	Mean [255-555] (mag)	Stddev	Mean [555-814] (mag)	Stddev
Nuclear	10	-0.90	0.54	0.38	0.24
Clump B	8	-1.72	0.30	0.00	0.17
Clump C	12	-1.70	0.23	0.32	0.19
No Region	18	-1.43	0.33	0.23	0.30

Thus, we have 2 possible interpretations for the Clump A nuclear region star cluster population: (1) the clusters suffer little reddening ( A_v < 0.2 everywhere) in this region, and some of the clusters are young ( $t_{{\rm age}} \sim 6$ Myr), some are intermediate (30 Myr  $< t_{{\rm age}} < 50$ Myr), and some are older ( $t_{{\rm age}} > 100$ Myr). Or, (2), there is significant and variable extinction (A_v = 0.1-1.0) in this region, and all of the clusters are $\sim$7 Myr. Additional information, such as high angular resolution spectra (e.g., Glazebrook et al. 1999), is needed to resolve this issue.

6.1.2 Clump B: The blue blob

From the color map we find that this region has little, if any, reddening, which should simplify our interpretation of the cluster colors. Looking at Fig. 10, we can see immediately that almost all the objects are consistent with ages <6 Myr. The colors for 2 of the clusters (Numbers 3, and 5) are 0.1-0.2 mag too blue in [555-814], which could indicate contamination from nebular emission lines. We can easily mimic this effect by adding emission line fluxes to the F555W model magnitude. The clusters then overlap the models at ages 3-6 Myr. Three of the clusters (Numbers 1, 4, and 6) are consistent with ages <4 Myr with a few tenths of a magnitude of visual extinction.

If we assume an age for the clusters, we can use the inferred absolute F555W magnitude to estimate the masses. The clusters range from -12 to -14.5, and assuming an age of 5 Myr, they span a range in mass of $5 \times 10^{4}$to $5 \times 10^{5}~M_{\odot}$.

6.1.3 Clump C

This region is complicated by a diffuse background in the F255W filter, indicating the presence of many faint F255W sources below our detection limits. This is likely to be the case in all regions, however here there is a lack of very massive clusters as seen in the nuclear region and Clump B, which would prevent the detection of such a background.

The Clump C region does not match the models; the [555-814] colors are too red for their [255-555] colors (see Fig. 10). However, Clump C is a strong H$\alpha$ source, so we expect it to be young. A possible explanation for the colors is that we are over-subtracting a strong nebular background. This would make the F555W magnitude too faint. If we assume this is the case, the [555-814] color gets bluer and the [255-555] gets redder, moving them onto the model track around 6 Myr. The spread in F555W magnitudes is comparable to that of Clump B, but almost a magnitude fainter. For an age of 6 Myr, we estimate the star cluster masses lie between $2 \times 10^{4}$ and $2 \times 10^{5}~M_{\odot}$.

6.1.4 Other F255W sources: "Clump F'' and the central cluster of Clump D

What was designated as Clump F by previous researchers, is clearly dominated by a single star cluster, number 16 of the "no-region'' clusters in Table 2. This object matches the model for a cluster with an age of 4-5 Myr, with perhaps a small over-subtraction of nebular emission. The sibling to this cluster, number 14 of the "no-region'' objects, is also consistent with an age of 4-5 Myr if we consider a contribution of 0.1 mag in F555W from nebular line emission. With this small nebular correction and the age assumption of 5 Myr, the clusters have masses of $3 \times 10^{5}~M_{\odot}$ and $1 \times 10^{5}~M_{\odot}$.

The central cluster of Clump D (Number 1) falls slightly below the model track for a cluster with an age of 9-13 Myr, and its companion (Number 2) is off a bit farther, although along the reddening vector. Assuming the photometry is reasonably accurate and the models are not in error, these clusters can move to this area of the two-color plot only by being very young, with ages of 1 Myr, and having more than 1 mag of visual extinction. However, they are unlikely to be younger than 10 Myr, considering the lack of H$\alpha$ emission (see Sect. 4). Also, we expect little to no reddening from the color map. If we consider solar metallicity instead of Z=0.008, the model track moves blue-ward in [255-555], eliminating any disagreement with the data. But it seems unlikely that the metallicity in this region should be greater than that of the nuclear region. Rather, it seems more probable that Clump D has a lower metallicity.

We also consider the possibility that the model tracks are in error. Allowing for the well-known uncertainty in red-supergiant stellar evolutionary tracks (Langer & Maeder 1995; Mayya 1997; Origlia et al. 1998; Leitherer et al. 1999), especially at metallicities less than solar, it seems possible that the [255-555] color may be $\sim$0.2 mag too red. With all these possible factors, we prefer the explanation of a 9-13 Myr cluster with little to no extinction.

An age of 9-13 Myr also explains something noteworthy in the H$\alpha$image. H$\alpha$ emission is not seen in the center of Clump D, but at the edges (NE and SE of the central cluster). Could this be evidence for propagating star formation? Clusters are detected in these regions in the F555W and F814W filters (presented in Sect. 6.2.4), and while their colors are not remarkably blue, they are consistent with ages of 7-8 Myr. This is interesting, although certainly not conclusive.

6.2 F555W and F814W

Now we turn to the set of clusters selected in the F555W and F814W filters, where we have detected 268 candidate star clusters with our selection criteria. Splitting this population into regions as we did previously, we find 20 Clump A star clusters, 14 clusters in Clump B, and 18 in Clump C. Here we also consider Clump D, which was too faint at 255, but in two filters has 11 members. Thus, there are 200 candidate star clusters, mostly relatively faint, which are not associated with a particular clump.

The color magnitude plots for the clusters separated by region are shown in Fig. 13, and the color magnitude diagram for the entire cluster sample is shown in Fig. 14. The clusters span a wide range of colors, with no apparent trend with luminosity. At the bright end are clusters in Clump B and the nuclear region. The cluster population not associated with a particular region starts to appear about two magnitudes fainter than these. This means that the most massive clusters are born in the clumps, although cluster formation is occurring throughout the galaxy.

Figure 12: Histogram for the 268 clusters in the final F555W and F814W sample, transformed to Johnson V and I. The mean color of this distribution is 0.42, with a dispersion of 0.29.

Figure 13: Magnitude-color plots for the clusters separated by region. The overplotted model track is for an instantaneous burst of $10^{6}~M_{\odot}$, with ages of 1, 5, 10, 20, 50, 100, and 500 Myr marked with asterisks. A purely stellar emission model is used, i.e. nebular contributions are omitted (see Sect. 4.1). The track shifts down by 2.5 mag for $10^{5}~M_{\odot}$, with no change in the F555W-F814W color (see Fig. 1).

For comparison with studies of other star-forming galaxies, we convert our [555-814] colors to standard Johnson V and I using the transformations found in H95b. The histogram of cluster colors is shown in Fig. 12, using a bin size of 0.15 mag. Mean V-I colors are shown in Table 4, along with the mean [555-814] colors.

The mean color of this distribution, 0.42, is blue, and comparable to the young cluster population in the Antennae (0.4, age < 30 Myr, Whitmore et al. 1999), the blue cluster population of NGC 7252 (0.4, age $\sim 30$ Myr, Miller et al. 1997), and the cluster population of the 4 Myr starburst in NGC 1741 (0.4, Johnson et al. 1999). This suggests a similarly young age for the cluster population presented here. It is clearly younger than the cluster population of NGC 3921 (0.59, age $\sim 250$ Myr, Schweizer et al. 1996).

6.2.1 Nuclear

The mean color of this region is redder than the cluster population as a whole. Looking only at the color map, we would suspect this to be the case due to dust extinction. From Fig. 13, we can see this is indeed what is happening. There is a pronounced color-luminosity trend, where the faintest clusters are the reddest ([555-814] $\sim$ 0.65). There are two possible causes for a color-luminosity trend, age and reddening. We can rule out an aging effect since the [555-814] color of a cluster is nearly constant as it ages. Therefore, the observed trend must be due to dust extinction. There are three important exceptions to this trend; the three brightest clusters have red colors ([555-814] > 0.5). Either these are old and very massive clusters, or they are very young and have yet to destroy their dusty natal cocoons. This can only be determined from spectroscopy, or, perhaps from high angular resolution images in emission lines that would reveal connections to the surrounding ionized gas.

Figure 14: F555W magnitude and [555-814] magnitude color plot for the 268 two-filter selected clusters. The overplotted model track is for an instantaneous burst of $10^{6}~M_{\odot}$, with ages of 1, 5, 10, 20, 50, 100, and 500 Myr marked with asterisks. The track shifts down by 2.5 mag for $10^{5}~M_{\odot}$, with no change in the [555-814] color (see Fig. 1).

6.2.2 Clump B: The blue blob

The color spread in this region is small, approximately 0.2 mag in [555-814]. However, there is a tendency for less luminous clusters to be redder, suggesting the effects of a small amount of reddening, requiring a correction for $\sim$A_v=0.3 mag. If we make this small correction to the faint clusters, the color of this region becomes nearly zero in [555-814], and the color spread becomes even smaller, $\sim$0.15 mag. Now we can say something about the formation timescale of this region. From a dynamical point of view, we would not expect a region of this size, radius $\sim$0.4 kpc, to have an age spread less than 3 Myr (using v = 150 km s^-1). But for this small range in color, the age spread cannot be more than 3 Myr, could be smaller, and is only compatible with the models if the entire region is younger than 6 Myr. Thus, we conclude that the clusters in Clump B are less than 6 Myr with age spread $\leq$3 Myr. With this age, cluster masses range from $5 \times 10^{5}$ to $5 \times 10^{3}~M_{\odot}$.

6.2.3 Clump C

The mean cluster color in this region is redder than that of Clump B, closer to that of Clump D, but with a larger color spread. Only for the brightest clusters does there appear to be a trend of color with luminosity. The mean color of this region suggests an age less than 50 Myr, although the colors of the three-filter sample suggests that at least some of the clusters are less than 6 Myr. As we saw for the three-filter sample, there may be complicating nebular effects in this region.

 

 

Table 4: Mean V-I and 555-814 cluster colors for the two-filter selected sample.

Region	Number of Members	Mean [V-I] (mag)	Mean [555-814] (mag)	Stddev (mag)
Nuclear	20	0.47	0.56	0.18
Clump B	14	0.07	0.17	0.17
Clump C	18	0.33	0.43	0.18
Clump D	11	0.40	0.49	0.13
All Clusters	268	0.42	0.52	0.29

6.2.4 Clump D

The clusters associated with Clump D all have roughly the same color, but a range of luminosities. This could mean that they have the same age and different masses, or a small range in mass, and ages <30 Myr with a spread. With a radius of approximately 200 pc, an age spread greater than 1.3 Myr is expected. This region is significantly fainter than the other star forming regions in this galaxy, with the brightest member at only 20.7 mag at F555W, or -12.7 in absolute magnitude. For the derived age of 9-13 Myr (as discussed in the previous subsection), we find a mass of $2 \times 10^{5}~M_{\odot}$.