Likelihood of Using Stellar Absorption Lines as Wavelength Markers for Spectra
                                
                          J Offiong
                                
  
     It is well known that the spectra of most stars, i.e., those on or not too far off the
  main sequence, resembles that of a blackbody.  In addition to the Planck curve, there are
  a number of spectral features caused by absorption due to the stellar constituents.  The
  absorption lines are often called Fraunhofer lines.
  
     Attached are spectra of some solar type stars. The data covers the wavelength
  range 3500 - 9000 angstroms.  Note that the designation V means that the stars are on the
  main sequence.  Also available in the Stellar Spectra directory on-line are spectra from
  stars from many more spectral types.  They are post script files and can be printed.  There
  is also an rtf file "classes.rtf" indicating some of the features in the spectra.  The stellar
  spectra and description of the features were found on the world wide web, at url
  http://zebu.uoregon.edu/spectra.html.
  
     Clearly evident in this spectrum are Calcium II H and K lines around 3900�, the
  G-band which originates from CH absorption around 4300�, Magnesium I at 5150� and
  Sodium D near 5800�.  The exact values of these quantities can be found in general
  literature (e.g., Allen: Astrophysical Quantities).
  
     The biggest difficulty will be the selection of lines.  Ideally, suitable lines will be
  both strong and isolated.  However, the currently proposed resolution of the instrument
  (about 50�) will limit the number of usable lines.  Fortunately, given the response of the
  OSIRIS instrument for these same lines, those listed above should prove strong and
  "wide" enough to be detectable on at least two to three pixels of the CCD.
  
     In order to use an absorption line as a wavelength marker, it will be necessary to
  determine the spectral type of the star from which the light came.  This will require the
  matching of a measured spectrum to that of a well known spectrum.  Traditionally, this
  has been done manually, however, recent advances in artificial neural networks has made
  automated spectral classification possible.  This must be investigated in the future.  Also
  to be determined is an appropriate "short lists" of usable absorption lines for different
  spectral types.
  
     In conclusion, it is appears that it will be possible to use absorption lines for
  wavelength markers.  Adequate features exist that are of appropriate strength and width
  in the required wavelength range.  Still, more precise work has yet to be done to
  determine exactly which lines will be detectable.
  
  Note:  Also available on-line is an rtf document by Raisa Lehtinen "raisa.rtf," which goes
  into greater detail on the use of  spectral structures in the calibration of the OSIRIS
  instrument.