The 10 and 20 
m observations and the results presented in this paper
were performed and obtained as part of the Infrared Space Observatory's
(ISO) calibration programme.
The ISO ground-based preparatory programme working group (ISO-GBPPwg), chaired by H.J. Habing, set up the goal in 1990 to characterise accurately in the infrared some 400 stars covering the whole sky, as well as a wide range of spectral type and brightness (Jourdain de Muizon & Habing 1991). We have thus obtained J, H, K, and L narrow-band photometry for all stars and for the brightest ones, spectroscopy and/or N and Q bolometry (van der Bliek et al. 1996; Hammersley et al., this work and in preparation).
The stellar effective temperatures (
) were determined using on one hand
the Infrared Flux Method (IRFM, Blackwell et al. 1990, 1995), and on the
other hand the V-K versus relationship (Di Benedetto 1993, 1995). Both
methods were combined with the surface gravity and metallicity (Cayrel de
Strobel et al. 1992) to predict stellar spectra up to about 50m from
the Kurucz stellar model atmosphere grids (Kurucz 1993).
The 10 and 20 m bolometry was originally planned to be used to
test the model atmospheres and detect stars with infrared
excesses (hence would be unsuitable as ISO IR standards). However, it turned
out that, due to the lack of reliable mid-infrared standards, the 10 and 20 m observations could not be accurately flux calibrated using the
standard methods and so serve the original purpose.
In Sect. 2 we describe the star selection and the observations. In Sect. 3 we present the data reduction, in particular tackling the delicate problem of determining a zero-point. We describe the use of the models and evaluate the contribution of all the possible sources of errors. In Sect. 4, we discuss the validity of using models to extrapolate mid-infrared magnitudes from near-infrared data, thus producing a new set of mid-infrared standards.