1994A&A...289..202M

Through the Infra Red Flux Method (IRFM), the effective temperature of a star is derived from the ratio R between the total integrated flux and an infrared monochromatic flux. Here we show that the metallicity of the stars has to be taken into account since the derived T_eff varies from 0.5% to 1.6%, according to [M/H] and to T_eff, i.e. from 65K to 230K. The influence of the superficial gravity is less than 0.5% as long as logg is larger than 0.35. If logg=0.30 the effect becomes as large as that of a metallicity equal to 2.5 or 3 times the solar one. The new heavily blanketed Kurucz's models lead to a systematic increase of the T_eff as compared to those obtained from oldest models through the same method, from 1.5% up to 2.5%, i.e. 100K to 290K. The T_eff obtained with ATLAS8 is lower by 30K to 120K than that given by ATLAS9. For stars which T_eff is between 6000K and 7000K, the T_eff derived when using ATLAS9 instead of the grid computed with the code MARCS from Gustafson & Bell is larger by more than 1.5%, typically from 50K to 100K. The effect of the model on the derived T_eff through the IRFM is illustrated in the case of eight dwarf or giant bright stars.