4 The open cluster Collinder 228

Collinder 228 was discovered by Collinder (1931) during a systematic search of open clusters in the Milky Way, and lies between us and the group formed by Trumpler 14, Trumpler 16 and $\eta$ Carinae (Smith et al. 2000). Therefore we expect the cluster to be dominated by back and foreground stars contamination. Differential reddening is also expected, since the cluster is surrounded by a large nebula (see Fig. 6).

Figure 7: CMDs for all stars in the region of Collinder 228. The dashed line indicates the limiting magnitude reached by Feinstein et al. (1976).

4.1 Previous results

Feinstein et al. (1976) reported on UBV photoelectric photometry of 99 stars in the region of Collinder 228. They found that the bulk of the cluster is located in front of the complex of Trumpler 14 and 16, at about 2.5 kpc from the Sun. They also pointed out that some stars in the field of Collinder 228 might be members of that complex and hence more distant than the cluster. They assign to Collinder 228 an age of $5 \times 10^{6}$ yrs. While the bulk of the stars is closer and has a mean reddening E(B-V)=0.30, the stars lying beyond the cluster have a larger reddening $E(B-V)\approx 0.50$.

Tapia et al. (1988) obtained JHKL near-infrared photometry of 200 stars in the $\eta$ Carinae region, which comprises Trumpler 14, 15, 16 and Collinder 228 and 232. Out of these, 45 are in the field of Collinder 228. By analyzing the two color diagrams, the authors concluded that this cluster is $2.09\pm0.38$ kpc far from the Sun, and hence closer to us than the bulk of $\eta$ Carinae region population. Moreover they found that the mean reddening is $E(B-V)=0.64\pm 0.26$, much larger than the value previously reported by Feinstein et al. (1976).

Finally, a radial velocity survey has been conducted by Levato et al. (1990), who suggested that 30% of the cluster stars are binaries.

Figure 8: Collinder 228 stars brighter than V=17 in the (B-I)-(B-V) plane.

4.2 The present study

We provide UBVRI photometry for about 1100 stars in a $3\hbox{$.\mkern-4mu^\prime$ }3 \times 6\hbox{$.\mkern-4mu^\prime$ }5$ region centered in Collinder 228, up to about V=21. The covered region is shown in Fig. 6.

The CMDs for all the measured stars are shown in Fig. 7 in the planes V-(B-V), V-(V-I) and V-(V-R). The MS extends from V=10 down to V=21. As for NGC 3114, the MS gets wider at increasing magnitude, and there is evidence of a secondary sequence on the red side of the MS, generated by the RG stars in the field.

 

 

Table 5: Photometry of the stars in the field of the open cluster Collinder 228 in common with Feinstein et al. (1976). The suffix CP refers to the present study, whereas FMF indicates Feinstein et al. (1976) photometry.

ID	FMF	Name	$V_{\rm CP}$	$(B-V)_{\rm CP}$	$(U-B)_{\rm CP}$	$V_{\rm FMF}$	$(B-V)_{\rm FMF}$	$(U-B)_{\rm FMF}$
1	15	HD 305544	8.656	0.694	0.192	8.59	0.66	0.08
2	28	HD 305543	9.778	0.073	-0.768	9.74	0.05	-0.77
3	29	HD 305451	10.284	0.112	-0.265	10.21	0.07	-0.36
4	30		10.801	0.109	-0.653	10.80	0.05	-0.69
5	49		11.125	0.200	-0.220	11.20	0.26	-0.34

We have 5 stars in common with Feinstein et al. (1976), which are listed in Table 5. The mean differences turn out to be:

$$V_{\rm CP} - V_{\rm FMF} = 0.021\pm0.054$$

$$(B-V)_{\rm CP} - (B-V)_{\rm FMF} = 0.019\pm0.041$$

$$(U-B)_{\rm CP} -(U-B)_{\rm FMF} = 0.073\pm 0.046 ,$$

where CP indicates our photometry, and FMF stands for Feinstein et al. (1976). Taking into account the different techniques used in extracting the photometry, the agreement is very good both for magnitude and colors.

Figure 9: Two color diagram for the stars in the field of Collinder 228 brighter than V=17. The arrow indicates the reddening vector. The solid line is the empirical un-reddened ZAMS from Schmidt-Kaler (1982), whereas the dashed line is the same ZAMS, but shifted by E(B-V)=0.30. See text for details.

4.3 Reddening

In order to obtain a rough estimate of the cluster mean reddening we consider the distribution of the stars brighter than V=17 in the (B-I) vs. (B-V) plane (see Fig. 8). The selection in magnitude is done to limit the field stars contamination. By applying the same technique described in Sect. 3.3, we find that the bulk of stars have $E(B-V)=0.40\pm 0.20$. The uncertainty is due to the scatter of the stars in this plane, and indicates the presence of stars with different reddening, in agreement with Feinstein et al. (1976) findings.

We use the (U-B)-(B-V) diagram for all the stars brighter than V=17 to separate cluster candidate members (see Fig. 9). The solid line in this plot represents the empirical un-reddened ZAMS from Schmidt-Kaler (1982), whereas the dashed line is same ZAMS, but shifted by E(B-V)=0.30.

Two distinct populations seem to exist. With filled circles we plotted all the stars having $E(B-V)=0.30\,\pm\, 0.05$, and we shall refer to them as to cluster candidate members. A second population is defined by stars having larger reddening and is plotted with open triangles. All these stars are probably just field stars.

These findings confirm Feinstein et al. (1976) results.

Figure 10: Reddening corrected CMDs for the candidate member stars in the field of Collinder 228. Over-imposed are solar abundance isochrones for an age of $8\times 10^{6}$ yrs. See text for details.

4.4 Age and distance

In Fig. 10 we plot the reddening corrected CMDs for the stars of Collinder 228 having $E(B-V)=0.30\pm 0.05$ (i.e. the candidate members). They actually seem to form a tight sequence, confirming our suggestion that they are good candidate members. Over-imposed is a theoretical solar metallicity isochrone (solid line) from Girardi et al. (2000) for an age of $8\times 10^{6}$ yrs. The same isochrone has been shifted by 0.75 mag (dashed line) to have an idea of the MS broadening due to unresolved binaries. It is well known, in fact, that binary stars define a sequence 0.75 mag brighter than the single stars MS. This permits us to suggest that five stars (indicated by open circles) are probably non members, and that the four stars which lie close to the binary sequence are probably unresolved binaries.

All the stars fainter than V = 10.5 in these plots have not been measured by Feinstein et al. (1976), and hence we provide 11 new candidate members. As a by-product, we infer an apparent distance modulus ( $m-M)=12.55\pm 0.25$, which, once corrected for extinction, provides a distance of $1.9\pm 0.2$ kpc, in agreement both with Feinstein et al. (1976) and with Tapia et al. (1988).

This corroborates the conclusion that Collinder 228 is closer to us than the Carina nebula complex.

In conclusion, in the observed region we identified 14 members, 3 in common with Feinstein et al. (1976) and 11 new, whose properties are summarized in Table 6. Out of these, 4 are probably binaries. Unfortunately, there is no overlap between our suggested binaries and the study of Levato et al. (1990). Only two stars are in common: HD 305543, which is a member and HD 305451, which probably is a field star. Nevertheless our result, which comes from photometry, although based on a small sample confirms their spectroscopic investigation that about 30% of the member stars are probably unresolved binaries.

Figure 11: DSS map of a region around vdB-Hagen 99. The box confines the field covered by our photometry.

 

 

Table 6: List of the new candidate members of Collinder 228 obtained in the present study.

ID	X	Y	V	(B-V)	(U-B)	(V-R)	(R-I)
4	-80.27	410.73	10.801	0.109	-0.653	0.043	0.081
5	73.19	205.14	11.025	0.200	-0.220	0.084	0.174
8	283.84	268.06	12.012	0.153	-0.421	0.072	0.140
9	-228.80	332.70	12.002	0.171	-0.493	0.071	0.118
12	-249.81	193.19	12.275	0.293	-0.206	0.154	0.198
13	109.62	331.36	12.190	0.191	-0.114	0.070	-0.226
14	-386.07	261.02	12.187	0.319	0.354	0.100	0.138
15	-227.63	41.27	12.599	0.354	0.383	0.118	0.223
18	268.82	85.25	13.269	0.270	0.145	0.160	0.227
19	-277.53	447.27	13.218	0.360	0.371	0.147	0.206
20	484.68	411.66	13.369	0.286	0.135	0.168	0.214