Astron. Astrophys. 362, 959-967 (2000)

3. Analysis and discussion

3.1. Photometric analysis

The observed () and () CMDs are shown in Figs. 2a and 2b, respectively, wherein the reddening vector for E(B-V) = 0.5, assuming A_v=3.0E(B-V) and E(V-I)/E(B-V) = 1.25 (Walker 1985, Straizys 1990), is also drawn. All CMDs exhibit star sequences with overall properties and some specific peculiarities. A common feature is that the upper part of star sequences is not parallel to the reddening vector but rather follows the envelope of Main Sequences (MSs) of relatively young open clusters. This feature is less obvious for BH 245 whose () diagram only shows two very sparce groups of stars, which become stellar sequences in the () plane. Field star sequences have a lower envelope with a smaller curvature than that of the Zero Age Main Sequence (ZAMS). This envelope does not depend on the space star density but on the parameters of the interstellar extinction, namely R = A_v/E(B-V) (Burki & Maeder 1973). The redder star sequence in the field of BH 245 seems to correspond to the object because it has a wider magnitude range in the () plane (with fainter and brighter stars) and it is not as faint as the bluer star group in the () diagram. The redder sequence is also not only less populous than the bluer one, in good agreement with the very poor star richness quoted by vdBH75, but also more tilted with respect to the direction of the reddening vector. The remarkable difference in the number of stars between both CMDs is mainly due to reddening effects.

Fig. 2. a (V,) and (V,) diagrams of stars in the fields of Ruprecht 119 (top) and NGC 6318 (bottom); b  (V,) and (V,) diagrams of stars in the fields of BH 245. The direction and size of the reddening vector for E(B-V)=0.5 mag, assuming Av=3.0E(B-V) and E(V-I)=1.25E(B-V), are also shown.

The CMDs of Ruprecht 119 present a long MS which extends between the eleventh down to the eighteenth magnitude, largely superseding the photometry of MV73 which reaches mag. This is probably the main reason why Ruprecht 119 was considered not to be a cluster by MV73. The brighter portion of the cluster MS is particularly well defined, while there are two apparent gaps in the () diagram at () (13.5, 0.6) and (16.0, 0.8). These gaps may be indicators either of differential reddening or of evolutionary effects (see, e.g. Canterna et al. 1979). Differential reddening could be caused by the presence of dust within the cluster because the bluest point of the MS is not as red as it would be expected if the cluster suffered from a high interstellar absorption. Note also that the MS in the () diagram is not as broad as that in the () diagram. Anyway, both hypotheses for the gaps origin indicate that Ruprecht 119 should be a young cluster. The CMDs of NGC 6318 also reveal the presence of a well-defined MS, the () diagram being at least one magnitude deeper than the () diagram, which is essentially due to reddening effects. Both CMDs exhibit no significant star field contamination.

With the aim of finding out whether stars in the field of BH 245 are distributed in the CMDs along a sequence or randomly, we first determined the cluster centre and then carried out circular extractions of stars located within 50 (22".5) and 100 (45".0) pixels. The cluster centre was determined from star density profiles built in the X and Y directions using all the measured stars, and from the geometrical centre taken from the corresponding finding chart. The adopted weighted value for the position of the cluster centre turned out to be () = (250, 260). For completeness purposes we repeated the same analysis for Ruprecht 119 and NGC 6318, the resulting cluster centres being (196, 212) and (219, 300), respectively. Figs. 3a and 3b show the extracted CMDs, wherein filled circles and crosses represent stars within 50 and 100 pixels from their cluster centres, respectively. The smallest circular extraction allows us to know which are the CMD regions where the fiducial cluster sequences are located, while the largest circular extractions represent a compromise between minimizing the unavoidable field star contamination and maximizing the number of cluster stars. As can be seen, most of the stars located within 50 pixels from the centre of BH 245 are distributed along the red sequence, in good agrement with the description of this object given by vdBH75. Based on this result, we conclude that BH 245 is a genuine open cluster which could be few Myrs old as judged from the steepness of its MS. Circular extractions for Ruprecht 119 and NGC 6318 allowed us only to trace the fiducial sequences for the cluster core regions, since their CMDs are clearly dominated by their MS stars. Particularly, circular extractions of Ruprecht 119's CMDs did not result in a clear tracing of its MS because the object is probably more extended than the total field covered by the CCD (cluster angular diameter = 6´.2, Alter et al., 1970).

Fig. 3. a Colour-magnitude diagrams of stars in the fields of Ruprecht 119 (top) and NGC 6318 (bottom): all measured stars (dots), circular extraction for r22".5 (filled circles) and r45".0 (crosses) are superimposed; b Colour-magnitude diagrams of stars in the fields of BH 245: symbols same as a .

We then determined the cluster reddening values, distances and ages by fitting the cluster MSs in the () and () diagrams to the ZAMS of Schmidt-Kaler (1982) and Piatti et al. (1998b), respectively. Bearing in mind that MSs alone can satisfactorily be fitted using different combinations of reddening and distance modulus, especially for young open clusters, we decided to use as reference the E(B-V) colour excesses and ages provided by the spectroscopic analysis (see Sect. 3.2). Firstly, we used spectrocopic E(B-V) colour excesses and the Schmidt-Kaler's ZAMS to derive distance moduli, which in turn were used to determine E(V-I) colour excesses by fitting MSs to the Piatti et al.'s ZAMS, the ratio E(V-I)/E(B-V)=1.25 being used. Ages were also estimated by matching the cluster MSs to the empirical isochrones of Piatti et al. (1998b, Fig. 3). Secondly, using spectrocopic ages and the empirical isochrones of Piatti et al., we matched cluster MSs to the corresponding isochrone curve and derived E(V-I) colour excesses and distance moduli. Thirdly, we use the spectroscopic ages and reddenings simultaneously to enter in the age-calibrated vs diagram to derive distance moduli and then E(B-V) colour excesses. Finally, the adopted fundamental parameters were obtained by properly combining all of the resulting values. Fig. 4 shows the empirical isochrones of Piatti et al. with the cluster MSs superimposed, while Table 3 gives in succession the E(B-V) and E(V-I) colour excesses, and the distance modulus and age for each cluster together with their corresponding errors. The last values are based on the uncertainties arising from the matching of the MSs to the ZAMSs and the age-calibrated vs diagram. Note that we only provide an age range because of the intrinsic MS's scatter. On the other hand, using the three independent determinations for both colour excesses we derived a ratio E(V-I)/E(B-V)=1.160.07, which indicates that the interstellar absorption in the direction to the clusters approximately follows the normal extinction law.

Fig. 4. Mv vs (V-I)o diagrams with the isochrones of Piatti et al. (1998b) superimposed. Symbols are as in Fig. 3.

Table 3. Cluster fundamental parameters derived from the CMDs analysis.

3.2. Spectroscopic analysis

Fig. 5 shows the observed integrated spectra for the cluster sample normalized at = 6000 Å for comparison purposes. Each spectrum affected by reddening shows the combined stellar content contributing to the cluster integrated light, wherefrom each cluster age can be inferred. For the present cluster sample the different shapes and continuum slopes essentialy reflect the result of reddening effects, since the clusters are of similar ages (see Sect. 3.1). In particular, the integrated spectrum of Ruprecht 119 presents emission lines. The nebular lines and and evidence of in the red wing of H can be seen in Fig. 5. H itself is very strong and it probably arises from a nebular component combined to extended stellar atmospheres' emission. Nebular lines in the range 5-30 Myr can occur in the integrated cluster spectra, as observed by Santos et al. (1995) in the Magellanic Clouds clusters. They are related to winds and supernova remnants, residual emission related to fossil HII regions, or diffuse emission in HII/OB association complexes, or finally projection effects. Stars with extended atmospheres like Be also occur in a similar age range (Mermilliod 1981a,b) and they show up in the integrated spectra of Magellanic Clouds clusters (Bica et al. 1990). On the other hand, the flux value of BH 245's spectrum in the near-infrared ( 9000 Å) relative to that at the normalization point ( 6000 Å) results in 2.5-3.0 times higher than those of Ruprecht 119 and NGC 6318, which suggests that the cluster suffers from a significantly stronger interstellar absorption. Otherwise, if no information about the cluster age is available, a noticeable high reddening can simply be inferred from the pronounced positive slope of its spectrum. The cluster flux is roughly constant and very low in the blue range up to 5000 Å where it undergoes an abrupt change to increasing values until reaches a value three times higher at the end of its observed spectrum. For this reason, it is very difficult to recognize any spectral feature in the blue-visible range; the continuum slope being the most relevant characteristic along the spectrum. For 8000 Å the spectrum also shows some important residuals introduced during the subtraction procedure of atmospheric emission bands.

Fig. 5. Observed integrated spectra in absolute units normalized to = 1 at = 6000 Å.

Integrated spectra were used to derive reddening values and ages of the cluster sample according to the precepts pointed out by Bica & Alloin (1986a, 1987). They studied integrated spectra in the visible and near-IR ranges of Galactic open and globular clusters, as well as Magellanic Clouds clusters. They examined the behaviour of metallic and Balmer line equivalent widths, as well as the continuum energy distribution in the spectral range 3700-10000 Å. They also generated a library of template cluster spectra with well-known properties. The fundamental cluster parameters were determined using the SPEED spectral analysis software (Schmidt 1988) at the Astronomical Observatory of Córdoba. First, we estimated cluster ages from equivalent widths (Ws) of Balmer lines in absorption by interpolating these values in the age calibration of Bica & Alloin (1986b), each W providing an age estimate. The ages derived from this method are reddening independent. Ws of absorption lines were measured according to the spectral windows and continuum tracings as defined in Bica & Alloin (1986a, 1987). Errors affecting the derived Ws were estimated from different measurements of the Balmer lines using high and low continuum tracings in order to take into account the spectral noise. In the case of BH 245 we could only estimate Ws for H and H because of the low spectral signal at these wavelengths. Table 4 lists the Ws of the Balmer lines and the resulting cluster ages obtained from the average of the independent values.

Table 4. Cluster fundamental parameters derived from the spectroscopic analysis.

We then used the ages provided by the Balmer lines to select an appropriate set of template spectra to derive cluster reddening values. The suitable template resulted to be the YA spectrum (t 3-6 Myr) from Santos et al. (1995), which basically corresponds to the NGC 2362 age group of Mermilliod (1981a,b). We also included in the list the YB (6-9 Myr) and YC (12-40 Myr) templates (Santos et al. 1995). These templates have flatter continua than YA due to the presence of luminous evolved stars, for example supergiants, which is well-documented spectroscopically in the near-IR (Bica et al. 1990). YA, YB and YC spectra cover the spectral range between 3500 and 6000 Å where reddening effects are more noticeable. We note that not much additional reddening information would be provided if near-IR templates with similar age resolution were available. The E(B-V) colour excess of each cluster was derived by matching the observed spectrum to that of the template that most resembles it, thus making use of the full spectral distribution. The age of these templates were also taken as additional independent estimates of cluster ages. Fig. 6 shows the reddening corrected spectra of the cluster sample compared to the YA spectrum, which best matches both continuum distribution and Balmer lines simultaneously. In the figure we applied arbitrary constant offsets to the cluster spectra for comparison purposes. The general appearance of cluster spectra is very similar to that of YA, although the larger the colour excess, the noisier the spectrum for 4000 Å. Balmer line intensities look similarly deep as expected because template ages were chosen from ages derived using Balmer line Ws. Note also that cluster continuum slopes ( 4000 Å) follow the trend of YA continuum quite well. If instead of YA template we had used YB, which is only 3 Myr older, then the colour excesses would have been 40% lower on average. This fact shows that the matching template technique is very sensitive to both age and reddening determinations, the typical E(B-V) error being 0.05 mag. Likewise, it must be noted that template spectra are the average of integrated spectra of several different clusters so that they represent the stellar population of clusters within a limited age range. On the other hand, a cluster spectrum reflects the behaviour of the combined light coming from its members, and is consequently more dependent on stochastic effects arising from the small number of stars, as it is the case in most young open clusters in the Galaxy. The adopted template age and reddening determinations are listed in Columns 7 and 8 of Table 4.

Fig. 6. Reddening-corrected integrated spectra compared with the template spectrum YA. Cluster spectra were shifted by arbitrary constants for comparison purposes.

3.3. Cluster parameters

We averaged cluster parameters derived from photometric and spectroscopic methods due to the very good agreement found between values obtained from both techniques. Table 5 lists the resulting final cluster properties. With the aim of conservatively estimating the interstellar absorption in front of the cluster, we decided to adopt the lowest value of both estimates, which closely corresponds to the mean value. Using these E(B-V) colour excesses and the apparent distance moduli of Table 3 we computed cluster distances and their corresponding uncertainties with the expression: (d) = 0.46[(V-M_v) + 3(E(B-V))]d, where (V-M_v) and (E(B-V)) represent the estimated errors in V-M_v and E(B-V), respectively. For the E(V-I) colour excesses, we directly adopted the photometric values. Finally, cluster ages were determined averaging with the same weight the youngest and oldest ages obtained from Balmer line Ws, template matching and isochrone fitting methods. The uncertainties of the adopted ages correspond to half of the difference between the lowest and highest age values.

Table 5. Adopted cluster fundamental parameters.

From the resulting ages we conclude that clusters have left the HII region phase which as a rule persists up to about 5 Myr. The integrated light of the clusters appears to be dominated by upper MS stars. No red supergiants are present in the cluster spectra and CMD's; however, their derived ages are compatible with their occurrence. These underpopulated clusters must be suffering stochastic effects which prevent one to fully observe all products of massive star evolution expected for their ages. In addition, no evidence of substantial internal reddening was found as to infer current star formation processes. The emission of H in the integrated spectrum of Ruprecht 119 might be an indicator of the presence of Be stars in the cluster, which could also be responsible for the gaps seen in the () diagram. BH 245 is very much reddened and is distant only 1 kpc from the Sun. A more extreme case is that of the young open cluster Westerlund 1 (Piatti et al. 1998c) which is possibly the most reddened open cluster optically observable, also located at about 1.0 kpc. Nearby individual dark clouds and complexes are known to exist in several disk directions (Cambrésy 1999), and therefore it would be important to explore in more detail the directions towards both BH 245 (Sagittarius, projected close to the Galactic centre), and Westerlund 1 in Ara, close to the Scorpius borderline.

© European Southern Observatory (ESO) 2000

Online publication: October 30, 2000
helpdesk.link@springer.de