2. Observations and results
2.1. H2O radio observations
The observations were performed on August 19, 20 of 1997, using the 64 m radio telescope at Parkes, Australia. We observed the rotational transition of ortho-H2O at a rest frequency of 22.235 GHz, using the 1.3 cm receiver plus autocorrelator backend. The 64 MHz band width and 1024 channels centred at 22.21 GHz yield a velocity coverage of 860 km s-1 at 0.84 km schannel-1. Using the Dual Circular feed we simultaneously obtained spectra in left and right circular polarization. The beam FWHM is , the system temperature is typically 110 K, and the conversion factor from antenna temperature to flux density is 6.3 Jy K-1. The nearby sky was measured every two minutes, resulting in very flat baselines that required only a very shallow second-order polynomial to be subtracted.
We observed the LMC RSG IRAS04553-6825 for six hours on-source integration. No difference was found between the spectra obtained at either polarization, which we then averaged. The final 22-GHz spectrum of IRAS04553-6825 is presented in Fig. 1, with flux densities in mJy and heliocentric velocities in km s-1. The measured rms (1 ) is only 5.5 mJy.
We detected a single peak of H2O maser emission with a peak flux density of 68 mJy (with 20% calibration accuracy), corresponding to a 12 detection. The peak is centred at a heliocentric velocity of 277 km s-1, and has a FWHM of 1.7 km s-1. The integrated flux of the emission is 0.17 Jy km s-1.
2.2. CaII triplet echelle spectroscopy
We used the 3.5 m New Technology Telescope (NTT) at the European Southern Observatory (ESO) at La Silla, Chile, on January 7, 1996, with the ESO Multi-Mode Instrument (EMMI), to obtain an echelle spectrum of IRAS04553-6825. Grating #14 and grism #4 as cross disperser were used, yielding a spectral coverage of 6000-9000 Å. The slit width and length were and , respectively. The integration time was one hour.
The data were reduced in the normal way using the Munich Interactive Data Analysis Software (MIDAS) package. The wavelength calibration was done by taking a ThAr lamp spectrum in conditions identical to the spectrum of IRAS04553-6825. The measured spectral resolving power is .
The equivalent width of the Ca II triplet lines at 8498, 8542, and 8662 Å measures the surface gravity in giants and supergiants, if the metallicity is known. Reversely, if an estimate of the surface gravity is available from knowledge of spectral type and luminosity, the Ca II lines can be used to infer the metallicity.
The echelle spectrum of IRAS04553-6825 around the Ca II triplet is presented in Fig. 2, on an arbitrary flux scale. The equivalent widths of the two strongest components are measured as Å and Å. Their sum is Å.
The Ca II absorption is maximum at a heliocentric velocity of 300 km s-1, corresponding to a redshift with respect to the stellar velocity of a little less than the outflow velocity in the outer parts of the CSE. This is explained by scattering of the photospheric spectrum by an extended, expanding dust shell. The measured equivalent widths of the Ca II lines are not affected by the scattering (see Romanik & Leung 1981, and references therein).
For IRAS04553-6825 we adopt a bolometric luminosity of , and an effective temperature between 2500 and 3000 K (spectral type: M7 supergiant). The surface gravity must then be about cm s-2. If we then compare the measured equivalent width of the Ca II triplet with the results from García-Vargas et al. (1998), we find that consistency is only achieved if the metallicity of IRAS04553-6825 is a factor two to three lower than solar. This result is not very sensitive to the values of the adopted stellar parameters, and hence we conclude that IRAS04553-6825 has a typical LMC metallicity ().
© European Southern Observatory (ESO) 1998
Online publication: August 6, 1998