next previous
Up: Spectral classification of unidentified


2 Observation and classification of unidentified sources

It would be unthinkably inefficient to obtain the individual spectra of the unidentified IRAS sources (in view of their extremely large number) and to classify them. On the other hand, objective-prism plates taken with a Schmidt Camera enable one to classify the spectra of all objects covered in the field in a single exposure. The wavelength region most appropriate for obtaining spectral information on unidentified IRAS sources is the photographic infrared. In this region, between 0.7 ${\mu}$m and 0.9 ${\mu}$m, strong diatomic bands dominate the spectra of cool stars: TiO and VO in the M stars, CN in the cooler (N-type) carbon stars, and LaO in the pure S stars. These features can be distinguished readily in objective prism spectra at dispersions as low as 3400 Å ${\rm mm^{-1}}$ at the telluric A-band and have been used to discover and classify cool stars since Nassau and colleagues pioneered in the use of Kodak I-N emulsion with objective prisms in the late 1940's (Nassau & van Albada 1949).

DJM (MacConnell 1992) collected I-N spectrum plates of dispersion 3400 Å ${\rm mm^{-1}}$ with the Curtis Schmidt Camera at CTIO; the deepest reach I $\sim$ 13.5. The plate scale is 96$.\!\!^{\prime\prime}$${\rm mm^{-1}}$, and each covers 5$\hbox{$^\circ$}\times 5\hbox{$^\circ$}$. They cover the galactic latitude belt $\pm7{\hbox{$^\circ$}}$ along the full southern galactic plane. In the years 1986-92, DJM classified about 14200 IRAS sources (identified as well as unidentified) in the R.A range $5^{\rm h}17^{\rm m}$ to $14^{\rm h}$ with $ F{_\nu}(12~{\mu}{\rm m}) \geq F{_\nu}(25~{\mu}{\rm m})$ falling in the area of the available plates. These constitute nearly 40$\%$ of all IRAS PSC sources of this type. The majority of these IRAS sources have cool stellar photospheres in the visible region, and the above colour criterion ensures that their CSE are warmer than 300 K. Stars earlier than type M3 have TiO bands too weak to be seen on these plates and so cannot be classified.

  
\begin{figure}
\includegraphics []{1490f1.eps}\end{figure} Figure 1: Reduced-scale version of a I-N objective prism spectrum plate obtained at a dispersion of 3400 Å ${\rm mm^{-1}}$ with the Curtis Schmidt Camera at CTIO, Chile. The deepest plates reach I $\sim$ 13.5. Each plate covers 5$\hbox{$^\circ$}\times 5\hbox{$^\circ$}$. The plate scale is $96\hbox{$.\!\!^{\prime\prime}$}6~{\rm mm}^{-1}$. The different symbols on the photographic print refer to IRAS sources with $ F{_\nu}(12~{\mu}{\rm m}) \geq 
F{\nu}(25~{\mu}{\rm m})$.The size of the symbol is proportional to $ F{_\nu}(12~{\mu}{\rm m})$. These symbols refer to sources as listed below: 1 = 1000, star = 1100, x = 1110, * = 1111, triangle = 1101, x with bar on top = 1010, x with bar top/bottom = 1011, infinity = 1001. The four numbers on the right of the sign "='' refer in the order to the flux density of these sources at 12, 25, 60 and 100 ${\mu}$m respectively, where "0'' means flux density is below the detection threshold and "1'' means flux density is above it

The plates were illuminated at a light table and examined at a magnification of 12X with a binocular microscope. Overlay plots to the plate scale and at the plate centers were generated for all plates at IPAC. Figure 1 shows a reduced-scale version of a plate on the appropriate IRAS plot. Stars from the SAO Catalog were also plotted to ensure the registration of plate with plot; small crosses denote positions of SAO stars. There was seldom any doubt as to which object on the plate corresponded to a given IRAS source even in the crowded galactic plane areas. The classifications and other data were stored on a microcomputer running a commercial program. Positions of the IRAS sources were measured with an x-y digitizer, and the measures were converted to equatorial co-ordinates using a specialized program on the PC which wrote the IRAS Name of the source into a database record and paused for entry of the classification and other data. These classifications are on file with the Astronomical Data Center (ADC) at the Goddard Space Flight Center, and most are contained in Version 2.1 of the IRAS PSC where they are designated as Catalog # 43. The sources classified in this way may be broken down into 10 groups, and the statistics based on a sample of 10500 classifications are listed in Table 1. The uncertainty in a type is $\pm1$ subclass.


  
Table 1: Statistics on the classification of IRAS Sources with $ F{_\nu}(12~{\mu}{\rm m}) \gt F{_\nu}(25~{\mu}{\rm m})$


\begin{tabular}
{lrlr}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
{\rm Typ...
 ...& Blank field:&6\\ \noalign{\smallskip}
\hline
\noalign{\smallskip}\end{tabular}


We examine here the data of only the unidentified IRAS sources which could be classified as O-rich stars (spectral type M) i) to search for correlations between spectral types and IRAS colours, and ii) to determine distributions of the various spectral classes in the two-colour diagrams.


next previous
Up: Spectral classification of unidentified

Copyright The European Southern Observatory (ESO)