5. Population ratios and radial gradients
Another useful application of the star counts in the different evolutionary phases is the calculation of the population ratios. In particular, here we can use the R-ratio (i.e the ratio of the number of the HB stars to the number of the RGs with bolometric magnitudes brigther than the HB magnitudes: ) proposed by Iben (1968) to derive the amount of the original helium in M 80. Table 6 presents the number of stars in the two branches, and , together with R values for three annuli at different distances from the cluster center and for the "complete" sample (stars with pix). The HB luminosity level of the RGB was fixed taking into account a differential bolometric correction between HB and RGB stars of 0.15 mag. The uncertainties shown in the Table take into account only the Poisson noise and neglect the errors coming from the uncertainty in the determination of which produces a variation of in R.
Table 6. Number of HB and RGB stars (, ), value of R for the specified annulus. The Table also shows the number of HB stars in the blue tail (), the number of stars brighter than blue tail stars (), and the ratios and (see text).
The "complete" sample gives a value of in full agreement with the value obtained by Buonanno et al. (1994) for M 3 () and, more generally, with those presented by Buzzoni et al. (1983) for a sample of 15 clusters. However, the quoted authors adopt as HB level the mean magnitude of stars in the RR Lyrae gap, while we use the lower envelope of the HB. Following the Buzzoni's approach we would have a slightly higher value of R: for the "complete" sample.
Adopting the calibration of Buzzoni et al. for the R-Y relation and the value from the "complete" sample, we have ; by using a more recent calibration (Bono et al. 1995), which involves the ZAHB level approach, we have (for Z=0.0005) , both in agreement with the generally accepted value of Y=0.23. Note that these determinations of Y must be considered as lower limits: our counts are incomplete at the limiting magnitude and we cannot exclude that the blue HB tail continues beyond the limit in our photometry. The problem of the field contamination on the blue HB tail has been investigated. Even if no conclusive results can be derived, we suggest that the blue tail stars belong to the cluster. First, by examining the field stars contribution in this region, as observed in a field located at arcmin from the center of the cluster, we found that no contamination is expected for . Second, some indication can be derived by using stellar population counts.
In this sense, even if affected by a large error, our determination of R allows to add further indication that the stars in the blue tail should be indeed genuine HB members. In fact, considering for example as "HB" only that part of the horizontal branch whose morphology is in common with that of M 3 (), we would have an amount of HB stars as given in column 5 of Table 6 () and a consequent value (i.e. ) which can hardly be accepted. Of course this method could be biased by any "non-evolutionary" phenomena acting in the way of modifying the ratios between stars in different evolutionary phases.
For the innermost sample ( pix) we derive a smaller value of R, which is probably due to incompleteness of our photometry in a more crowded region of the cluster. Nevertheless, the analysis of the outer (complete) annuli ( pix and pix) shows the occurrence of a radial gradient in the parameter R, from which one could infer that, moving toward the cluster center, an excess of HB stars and/or a depletion of RGB stars takes place (see Djorgovski & Piotto 1993 for very similar results in other high concentration globular clusters like M 15, M 30, and NGC 6397).
The same Table 6 presents the values for defined as the ratio , where is the number of stars in the blue tail (). Again, the values of for the two outer annuli reveal that the number of BT stars increases toward the cluster center. This appears a significant result, since those stars are fainter and more affected by incompleteness, more severe in the inner regions, than the brighter red horizontal branch (RHB) stars. A similar occurrence was found also in NGC1904 (Ferraro et al. 1992a). For a more significant comparison of the radial distribution of the stars in the different evolutionary phases, we have calculated the cumulative distribution. As shown in Fig. 5 the radial trend of the R and ratios are due to an increase in the number of the HB stars toward the cluster center (note that we have used only stars with , i.e. arcsec), and not to a dimise of RGB stars. In particular, if we keep the division in RHB () and blue horizontal branch (BHB) (), a Kolmogorov-Smirnov test shows that the BHB stars are significantly more concentrated (at 99.8% probability level) than the SGB stars (defined as stars within 3 sigma from the fiducial sub-giant branch and with ); the RHB stars are more concentrated, but this last result is less significant (Fig. 5, upper panels). The BHB stars are more concentrated (at the same significance level) even if we limit the BHB sample to or to : not only the result is significant, but we can exclude it is an artifact due to the inclusion of spurious detections, as it could happen extending the counts to the limit of the photometry. Neither the incompleteness can explain the radial trends shown in the upper panels of Fig. 5. On the other side, the RGB stars have exactly the same distribution as the SGB stars (Fig. 5, lower panel).
A similar result has been found in NGC 1851 (Saviane et al. 1998), though at a lower significance level, and also NGC 1904 shows the same effect.
Undoubtedly, the most interesting feature in the CMD of Fig. 1 is the prominent blue tail of the HB. Extended blue tails have been found in many other galactic GCs, and their nature and, overall, their origin is not well understood, yet (Sosin et al. 1997). As shown above, whatever the BHB stars are, they must be the natural descendant of the RGB stars, if we exclude anomalous helium content for M 80. However, while the RGB stars have the same distribution of the other fainter stars populating M 80, the BHB stars are significantly more concentrated towards the inner parts of the cluster.
It is hard to interpret this observational evidence, from the dynamical point of view. It cannot be mass segregation, as, interpreting the BHB stars as normal He burning stars, they must be less massive than RGB stars. On the other side, whatever the mechanism responsible of this stellar distribution is, it must be effective far from the cluster core. One possibility is that some of these blue tail HB stars are just stars that, due to some close encounters in the inner high density cores, have lost most of their envelope, and, at the same time, gained enough energy to be ejected from the core to the outer envelope. This mechanism could qualitatively explain the gradients in the BHB stars (the probability of gaining energy is lower for the higher energy corresponding to larger orbits). If this is the case, we would expect an anomalously high R ratio. The fact that the R ratio 'seems' normal can be explained by the fact that we do not count all the BHB stars, as the blue tail extends beyond the limit of our photometry. This attempt of explanation cannot be more than a simple speculation till we will be able to map the entire cluster from the inner core to the outer envelope to better track the stellar distribution. Calculation of the cross section for encounters able to strip part of the envelope and to leave the star with a higher energy would be most valuable. Whatever the interpretation of these results is, these features suggest the occurrence of local modifications in the stellar populations of the cluster in agreement with what is found in other dense and post-core-collapse globulars (see e.g. Djorgovski & Piotto 1993, Djorgovski et al. 1991, Castellani 1994, Fusi Pecci et al. 1993).
© European Southern Observatory (ESO) 1998
Online publication: June 26, 1998