The success of population and evolutionary synthesis calculations of the integrated light of clusters and galaxies critically depends on the availability of a suitable library of stellar spectral energy distributions (SEDs), which we shall henceforth call stellar library. Because of the complex nature of the subject, there are very many ways in which such calculations can contribute to the solution of any particular question relevant to the stellar populations and their evolution in clusters and galaxies. In previous studies, the particularities of these questions have largely determined the properties that the corresponding stellar library must have in order to be considered suitable for the purpose.
There is now a considerable arsenal of observed stellar libraries (for a recent compilation, see e.g. Leitherer et al. 1996), each of which has its particular resolution, coverage, and range of wavelengths as well as its particular coverage and range of stellar parameters - but which, even if taken in the aggregate, fall short by far of providing the uniform, homogeneous, and complete stellar library which is required now for a more systematic and penetrating exploitation of photometric population and evolutionary synthesis.
Ultimately for this purpose, what is needed is a uniform, homogeneous, and complete theoretical stellar library, providing SEDs in terms of physical parameters consistent with empirical calibrations at all accessible wavelengths. Thus, the above goal can be approached by merging existing grids of theoretical model-atmosphere spectra into the desired uniform and complete stellar library, and making it both homogeneous and realistic by empirical calibration.
A variant of this approach was first tried by Buser & Kurucz
(1992), who constructed a more complete theoretical stellar
library for O through K stars by merging the O-G-star grids of Kurucz
(1979a, b) with the grids of Gustafsson et al.
(1975), Bell et al. (1976), and Eriksson et
al. (1979) for F-K stars. In their paper, Buser & Kurucz solved
for uniformity and homogeneity by recomputing new late-type spectra for
Kurucz's (1979) standard grid of wavelengths and using the
Kurucz & Peytremann (1975) atomic opacity source tables.
The resulting hybrid library
has, indeed, significantly expanded the ranges of
stellar parameters and wavelengths for which synthetic
photometry can be obtained with useful systematic accuracy and
consistent with essential empirical effective temperature and metallicity
calibrations (Buser & Fenkart 1990; Buser & Kurucz
1992; Lejeune & Buser 1996).
In the meantime, Kurucz (1992, 1995) has provided a highly comprehensive library of theoretical stellar SEDs which is homogeneously based on the single extended grid of model atmospheres for O to late-K stars calculated from the latest version of his ATLAS code and using his recent multi-million atomic and molecular line lists. The new Kurucz grid - as we shall call it henceforth - indeed goes a long way toward the complete library matching the basic requirements imposed by synthetic photometry studies in population and evolutionary synthesis. As summarized in Table 1 (click here) below, SEDs are provided for uniform grids of wavelengths and stellar parameters with almost complete coverage of their observed ranges! These data have already been widely used by the astronomical community, and they will doubtlessly continue to prove an indispensable database for population and evolutionary synthesis work for years to come.
In the present work, we shall endeavor to provide yet another indispensable step toward a more complete stellar library by extending the new Kurucz grid to cooler temperatures. This extension is particularly important for the synthesis of old stellar populations, where cool giants and supergiants may contribute a considerable fraction of the total integrated light. Because model atmospheres and flux spectra for such stars - the M stars - have been specifically calculated by Bessell et al. (1989, 1991) and by Fluks et al. (1994), our task will mainly be to combine these with the new Kurucz grid by transformation to the same uniform set of wavelengths, and to submit the resulting library to extensive tests for its realism. In fact, as shall be shown below, the process will provide a complete grid of SEDs which is homogeneously and consistently calibrated against observed colors at most accessible wavelengths.
In Sect. 2, we shall briefly describe the different libraries used in this paper and the main problems that they pose to their unification. Because the spectra exhibit systematic differences both between their parent libraries and relative to observations, we set up, in Sect. 3, the basic empirical color-temperature relations to be used for uniformly calibrating the library spectra in a wide range of broad-band colors. This calibration process is driven by a computer algorithm developed and described in Sect. 4. The actual color-temperature relations obtained from the corrected library spectra are discussed in Sect. 5, and the final organization of the library grid is presented in Sect. 6. In the concluding Sect. 7, we summarize the present state of this work toward the intended standard library, and we briefly mention those necessary steps which are currently in process to this end.