2 Observation and data reduction

Spectroscopic data were obtained during two observing runs, in 1995 and in 1997, with the 2.16 m telescope at Beijing Astronomical Observatory. In 1995 the telescope was equipped with a Tektronix 1024 CCD mounted at an UNIVASAL spectrograph, whereas in 1997 with another Tektronix 1024 CCD at an OMR (Optomechanics Research Inc.) spectrograph. The pixel size of the two Tektronix CCDs were both 24 $\mu$m. For the UNIVASAL spectrograph a grating of 300 lines/mm were used, and for the OMR spectrograph three gratings of 300 lines/mm, 600 lines/mm and 1200 lines/mm were used. The corresponding dispersion and central wavelength are given in Table 1. A slit with a width of about 2'' was used during all the observations.

  

Table 1: Spectrographs and gratings

A total of 263 program stars, 140 from IC 1805, 46 from NGC 654 and 77 from NGC 6823, were observed. They are mainly proper motion cluster members (See Sect. 3.2). The magnitude limit is around $V=14^{\rm m}$. Spectra were taken with the 300 lines/mm grating for nearly all program stars. For the brighter stars spectra were also taken with the 600 lines/mm grating, and for the few brightest stars additional spectra were obtained with the 1200 lines/mm grating. The exposure times ranged from a few minutes to about an hour, depending on the magnitude of the object, the seeing, and the grating used. The resulting signal-to-noise (S/N) ratios usually exceeded 100. The observed cluster stars with their identifying numbers are listed in Tables 2, 3, and 4 for IC 1805, NGC 654 and NGC 6823 respectively.

  

Table 2: Spectral classification of stars in IC 1805

 

Table 2: continued

 

Table 2: continued

  

Table 3: Spectral classification of stars in NGC 654

  

Table 4: Spectral classification of stars in NGC 6823

 

Table 4: continued

We also observed a complete grid of MKK standards (up to 200) with the same instruments (telescope and spectrograph with same gratings) and under nearly identical observing conditions as our program stars. The S/N was kept comparable to that of the program stars. These standard stars are mainly from Keenan & McNeil (1976); Morgan et al. (1978) and Morgan & Keenan (1973). Reobserving MKK standards here is necessary for the spectral comparison and classification.

All the spectra were reduced using MIDAS. The raw CCD spectral images were corrected using average Bias and high S/N dome-flat frames. The dark current of the used CCDs was negligible. Sky subtraction was achieved by subtracting a third degree polynomial fitted to the spectrum in the spatial direction. This kind of sky subtraction proved to be effective in eliminating the emission from surrounding nebulosity. The sky-subtracted spectral images were then wavelength-calibrated using the comparison spectra (Fe-Ar or He-Ne-Ar lamps) and one-dimensional spectra were extracted. For the purpose of spectral classification all one-dimensional spectra were continuum-normalized instead of flux-calibrated. Data reduction for program stars and MKK standard stars were the same. Representative spectra of MKK standards are illustrated in Fig. 1 and Fig. 2. Spectra of all the program stars from the three clusters are shown in Fig. 3 through Fig. 8.

  

Figure 1: Representative spectra of MK standards taken with 600 lines/mm grating

  

Figure 2: Representative spectra of MK standards taken with 300 lines/mm grating