Astron. Astrophys. 351, 477-486 (1999)

4. Stellar population synthesis

To study the history of star formation, age, and ionization mechanism of BCGs, we applied the synthesis method of Schmitt et al. (1996) for a determination of the stellar population in their nuclear region. In this method, we start with a sample of star clusters, consisting of 3 clusters in the SMC, 12 in the LMC, 41 Galactic globular clusters and 3 rich compact Galactic open clusters, together with 4 HII regions (Bica & Alloin 1986). Then a grid of base components (comprising the values of the continuum at selected points and of the EWs of selected absorption lines) is constructed by interpolation and extrapolation in the plane of those quantities of the clusters of the sample. The grid consists of 34 points with ages at , and = 0.6, 0.3, 0.0, -0.5, -1.0,-1.5, -2.0, plus one point representing the HII region. The method, when applied to a given target, consists of adjusting the percentage contributions of the 35 base components by minimizing the difference between the resulting and measured EWs of the selected set of absorption lines. When all the resulting EWs reproduce those of the galaxy within allowed limits, we took this to be an acceptable solution. We then took all the acceptable solutions to form an average solution (Schmitt et al. 1996).

The computation can be performed in two ways: one way spans the whole plane (multi-minimization procedure, hereafter MMP), while the other is restricted to chemical evolutionary paths through the plane (direct combination procedure, hereafter DCP). We combine these two methods in our paper. We first use the MMP method to single out the main contributing components. And then, based on their resemblance to the whole-plane solution and on their reduced chi-square (), we select the best evolutionary path and use the DCP method to give the final result.

4.1. The results of MMP

A detailed analysis of the MMP method can be found in Schmidt et al. (1989). This method searches the vector space of resolutions generated by the entire 35 component basis, leading to a representative set of acceptable solutions to the synthesis problem. We tried various combinations of the 35 components until a good match between the equivalent widths of the synthetic and observed lines is obtained. An iterative optimization procedure was used, and each iteration alters the percentages of different components. The input parameters are the measured equivalent widths of the selected absorption lines, the continuum ratios and a set of trial values of between 0.0 and 1.0 at steps of 0.02. The result is expressed with the flux fractions at 5870 Å for each component. The flux fractions of different components for the 10 BCGs are shown in Table 4 (HIIR denotes the HII region component).

Table 4. Flux fractions (%) at 5870 Å for the whole plane solution for BCGs.

From the results of the MMP method, we find some obvious trends for all the BCGs. First, the dominant population is young () stellar clusters. Second, there is a small population of old () and high metallicity () globular clusters. Within old globular clusters low metallicity components contribute more than high metallicity components. Third, for the young and intermediate age () clusters, components with metallicities below or equal to the solar value make a large contribution. It shows that the stars in the BCGs have low metallicities, and this is consistent with our previous knowledge. Lastly, the younger components show a large dispersion in the plane, with no clear evolutionary paths. This could be due to the fact we have only used spectral data in the visible range. Additional data in the near ultraviolet and infrared will produce better-constrained solutions in the plane. Nevertheless, the presence of starbursts with is easily recognized from the MMP results.

4.2. The results of DCP

To reduce the dispersion in metallicity, all the population synthesis so far made assumed some arbitrarily chosen chemical evolution path in the plane. The improvement in the method of this paper is: we shall pick out from the MMP results, those components that contribute most importantly and use them to define paths of chemical evolution, thus reducing the degree of arbitrariness. The 3 bright BCGs appear to follow a path containing the 11 components along the time sequence at fixed metallicity , the metallicity sequence with , and the HII region. The 7 BCDGs follow a path containing the 12 components along the time sequence at fixed metallicity , the metallicity sequence with , and the HII region. In addition, we also tested other possible paths with different maximum metallicities; we found that the of the path selected from MMP is the smallest.

Table 5 reports the path solution. The numbers of acceptable solutions and corresponding reduced chi-squared () are also given. There is a great similarity between the path (DCP) and the whole plane (MMP) solution. Table 5a, 5g, 5j provide the results for the 3 bright BCGs (IC1586, NGC4194 and MRK499). We see that the younger components with age yr make an appreciable contribution, . The old globular clusters make even larger contributions. The intermediate-age components are small. The other tables in Table 5 provide the results for the 7 BCDGs. The dominant stellar components () are in the young age bins (). The old globular clusters have different values in different galaxies; in some, lower than , and in others as much as . The BCDGs differ from the bright BCGs in two obvious aspects. First, the intermediate age components in the BCDGs amount to , higher than in the bright BCGs. A peak occurs at about , which indicates that an enhanced star formation event occurred at that epoch. Second, the youngest components () in the BCDGs are much lower than in the bright BCGs; probably indicating that the star formation rate of BCDGs is lower now. The HII region component is a featureless continuum, which acts in the synthesis as a dilutor of absorption lines. From Table 5, we note that some BCDGs have large contributions from the young components indicating intense star formation, and small contributions from HII regions, which suggests that intense starbursts have converted most gas into stars.

Table 5. Flux fractions (%) at 5870 Å for the best path solution of BCGs.

© European Southern Observatory (ESO) 1999

Online publication: November 3, 1999
helpdesk.link@springer.de