The spectroscopic observations aimed at confirming the nature of the selected
candidates were carried out at the F/15 Cassegrain focus of the 1.93 m
telescope at the Observatoire de Haute-Provence (CNRS, France) during the
nights of April 8-15, 1997 and April 20-21, 27-29, 1998. The long-slit
spectrograph CARELEC [(Le-maitre etal. 1990)] was operated with a 150 lines
mm-1 grating leading to a dispersion of 26 nm mm-1 in the
wavelength range nm. The detector was a
Tektronix CCD with 27
m pixels and a rms read-out noise of
9.2 electrons. Throughout the observing runs the atmospheric seeing usually
ranged from
to
. Most of the observations were made
with a
slit width, which corresponds to a projected slit width on
the detector of 62
m or 2.3 pixels. With this set-up, the overall
instrumental resolution FWHM was 1.4 nm as measured on the He arc lines used
for the wavelength calibration. An exposure time of 3600 s was enough to yield
a signal-to-noise ratio larger than 10 in the spectra of the faintest
candidates (
).
The sub-dwarf O star HZ 44 [(Oke 1990)] was used as a spectro-photometric standard during both observing runs. Depending on the atmospheric stability of the night, the absolute flux calibration is uncertain at the 20 to 50% level. The data were reduced and analyzed independently by two of us, with the LONG-SLIT package implemented within MIDAS (CL), and with IRAF (DVG). After flat-field correction and wavelength calibration, the object signal was extracted using the Horne algorithm ([Horne 1986]). It ensures an optimal signal-to-noise ratio for weak spectra while it possibly slightly underestimates the emission-line strength in extended sources.
Copyright The European Southern Observatory (ESO)