Distant field blue horizontal branch stars and the mass of the Galaxy - I. Classification of halo A-type stars

This is the first in a series of three papers presenting a new calculation of the mass of the Galaxy based on radial velocities and distances measured for a sample of some 100 faint 16 < B < 20 field blue horizontal branch (BHB) stars. This study aims to reduce the uncertainty in the measured mass of the Galaxy by increasing the number of halo objects at Galactocentric distances r > 30 kpc with measured radial velocities by a factor of 5. Faint A-type stars in the Galactic halo have been identified from ub(j)r photometry in six UK Schmidt fields. These samples include field BHB stars as well as less luminous stars of main-sequence surface gravity, which are predominantly field blue stragglers. We obtain accurate charge-coupled device photometry and spectra to classify these stars. This paper describes our methods for separating out clean samples of BHB stars in a way that is efficient in terms of telescope time required. We use the high signal-to-noise (S/N) ratio spectra of A-type stars of Kinman, Suntzeff & Kraft (published in 1994), and their definitive spectrophotometric A classifications, to assess the reliability of two methods, and to quantify the S/N ratio requirements. First we revisit, refine and extend the hydrogen linewidth versus colour relation as a classifier (here called the D-0.15-colour method). The second method is new and compares the shapes of the Balmer lines. With this method (here called the scale-width-shape method) there is no need for colours or spectrophotometry. Using the equivalent width of the Call K line as an additional filter we find we can reproduce Kinman, Suntzeff & Kraft's A classifications with both methods. In a sample of stars with strong Balmer lines, equivalent width Hgamma > 13 Angstrom [equivalent to the colour range 0 less than or equal to (B - V)(0) less than or equal to 0.2], halo BHB stars can be separated from halo blue stragglers reliably. For the spectroscopy (i.e. both classification methods) the minimum required continuum S/N ratio is 15 Angstrom(-1). For the D-0.15-colour method (B -V)(0) colours accurate to 0.03 mag are needed.