It is well known that normal discs of spiral galaxies have a component which can be fitted to an exponential (Freeman, 1970) with a radial scale length h. There are two basic reasons to undertake a new determination of the parameter h in the Milky Way:
a) There is a very large scatter in the values found in the literature. For instance, Kent, Dame and Fanzio (1991) after reviewing previously reported values, found that they range from 1 to 5.5 kpc, and proposed an intermediate value of 3 kpc. van der Kruit (1986) obtained 5.5 kpc; Lewis and Freeman (1989) gave a value of 2.5 kpc, and so did Robin, Crezé and Mohan (1992), among others. This uncertainty is also too large in the NIR. Mikami and Ishida (1981) obtained a value of 1 kpc, Jones et al. (1981) 2 kpc, and Eaton et al. (1984) 3 kpc. Values of h in the NIR seem to be lower, but even this fact remains uncertain, despite its interest in the dynamic evolution of the disc.
b) Even if h is simply a fitting parameter for a distribution with frequent deviations from an exponential, it has became a natural radial length unit, very useful when comparing spiral galaxies with different sizes or unknown distances, and physically more significant than true length or angle units. The position of any morphological feature or typical lengths of disc phenomena are usually and hopefully expressed taking h as unit. Some of these features in our own galaxy may be considered normal or exceptional depending on the value of this normalizing length.
Probably the large scatter of data is a result of managing a large amount of data concerning different regions of the sky, which are difficult to analyse because of the contribution of many other complex components, such as bulge, star formation ring associated with the gas ring, bar, spiral arms, warp and so on. However, a proper choice of the wavelength and the region of the sky may render this problem a very simple one, which implies a more confident obtention of the parameter h.
This search specifically deals with, and aims at an improved determination of the radial scale length of the Milky Way disc.
We have chosen the K-band, in which extinction is very low. This also implies that the observed distribution closely resembles the true one. In addition, we have taken measurements at , rendering extinction negligible. We then surveyed a galactic longitude range from to . This practically eliminates the contribution of bulge, ring and bar (Hammersley et al., 1994) which take place for , and the contribution of the Local Arm, warp and truncation (Porcel et al., 1997), which take place for . The contribution of the other spiral arms (not the Local Arm) is minimized taking , as they remain closer to the plane. Another advantage of working in the K-band is that the observed stars in the 9-10 magnitude range are practically only K2-K5 III stars, through the surveyed sky zone. This fact was realised after taking into account the previous analyses of Ruelas-Mayorga (1991), Wainscoat et al. (1992), Calbet et al. (1995) and our own initial calculations (Porcel, 1997). The light coming from these stars is 80 % of the total light from all stars in this magnitude range. This greatly simplifies the problem as the luminosity function can be approximated by a Dirac's delta function.
Observations were carried out with the Carlos Sánchez 1.5 m telescope at the Teide Observatory, as part of the Two Micron Galactic Survey (TMGS) project (Garzón et al., 1993; Hammersley et al., 1994; Calbet et al., 1995). Interpolation techniques were used to obtain a regular mesh from the series of discrete scans (Porcel, 1997; Porcel et al. 1997).
© European Southern Observatory (ESO) 1998
Online publication: January 8, 1998