Infrared Beacons in the Universe - Red Supergiant Stars and the Chemical Composition of Galaxies

Red supergiant stars are very interesting astronomical objects; with future telescope facilities they can be used to probe the chemical composition of cosmic matter outside our own galaxy and even outside the so-called Local Volume comprising some 500 known galaxies. Recently, scientists at the Max-Planck Institute for Astrophysics have teamed up with researchers at the Hawaii Institute for Astronomy and the IoA Cambridge to analyse the physical properties of the photospheres of these red supergiants, focussing on the departures from local thermodynamic equilibrium. The primary goal is to explore how these complex phenomena influence spectral line formation, and thus the determination of the chemical composition from their observed spectra. The scientists found that estimates of element abundances could be wrong by more than a factor of 2, if these effect were neglected.

Fig. 1: An optical image of the young Milky Way stellar association Perseus OB-1, in which about 70 blue (BA-type) and 20 red (M-type) supergiants were found (appear red on the Figure).
Credit: F. Calvert and A. Block (NOAO, AURA, NSF, KPNO).

Fig. 2: A theoretical (green) and observed (black) infrared spectrum of the prototype RSG star, Betelgeuse. The advantage of the J-band over the H- and K-bands is obvious. The observed spectrum was taken from the IRTF spectral library (NASA IRTF, Mauna Kea; Cushing et al. 2005).

Fig. 3: Complete atomic model for the neutral titanium.

Fig. 4: The differences between NLTE and LTE Ti abundances derived from the RSG model spectra as a function of T_eff for three values of metallicity, [Fe/H] = -0.5 (white), 0 (blue), and +0.5 (red).

Fig. 5: Comparison of the observed (black) and theoretical NLTE (red) spectra for Betelgeuse.

Chemical composition is one of the key observable characteristics of star forming galaxies in the nearby and in the high redshift universe. So far, most of our information about their metal content (astrophysicists summarise all elements heavier than helium as "metals") has been obtained from the analysis of strong emission lines from H II regions, i.e. low-density clouds of partly ionised hydrogen gas. However, measurements of galaxy metallicities are then uncertain by a large factor because of the systematic uncertainties inherent in this 'strong-line' method. Furthermore, the method yields basically only the oxygen abundance, which is then taken as a placeholder for the overall metallicity. In this case, there is no information on abundance ratios, which can be a powerful diagnostic of the chemical enrichment history.

An alternative approach avoiding these weaknesses is the spectroscopic analysis of supergiant stars, the brightest stars in galaxies with luminosities up to one million times brighter than the Sun. Here, much progress has been made through the optical spectroscopy of blue supergiants in the Milky Way (Fig. 1) and few other galaxies of the Local group. For extragalactic astrophysics, however, red supergiants (RSGs) are more promising candidates. Their spectral energy distribution peaks in the infrared, where interstellar extinction is reduced. Particularly attractive for quantitative spectroscopy is the J-band (Fig. 2), which contains many isolated atomic lines. Spatial resolution in the infrared is also higher than in the optical and the advantage of instruments supported by adaptive optics can be fully exploited. RSGs are therefore ideal targets for spectroscopy with future telescope facilities, such as the Thirty Meter Telescope (TMT) and the European Extremely Large Telescope (E-ELT). The abundances of various chemical elements could be then directly measured out to distances of 70 Mpc, far beyond our Local Group of galaxies.

The analysis of RSG spectra, however, is a challenging task. One major complexity arises due to very low gravities (about 1000 times less than on Earth), stipulating departures from Local Thermodynamic Equilibrium (LTE) in their photospheres. Until now, it has not been possible to compute RSG spectra in non-LTE, as one would need detailed atomic data for constructing the non-LTE atomic models and to model the complex line blanketing, which is dominated by molecules. So far, the atomic data was of insufficient quality. Furthermore, while in LTE the line formation calculation are fairly simple, in the non-LTE case, accurate radiative intensities must be computed at all frequencies where radiative transitions occur in an atom. For iron and titanium this means for wavelengths from UV to far-IR. Such calculations were simply beyond the computational power so far.

The team of scientists has now constructed complete atomic models of neutral iron and titanium (Fig. 3) and performed, for the first time, non-LTE calculations of these atoms using model atmospheres representative of RSG stars. The inclusion of non-LTE effects changes the titanium abundances dramatically compared to LTE and with different correction factors for different temperatures and metallicities (Fig. 4). However, the non-LTE effects on the J-band iron lines are much smaller. Since the RSG metallicity is determined from many lines of different atomic species (iron, titanium, silicon and magnesium), the overall effect of non-LTE corrections may not be as large as for titanium alone. For example, the best non-LTE fit to the J-band spectra of Betelgeuse is shown in Fig. 5.

The next step therefore will be to model the other prominent species in the J-band spectral window, namely silicon and magnesium. These lines also contain important metallicity information, since together with Ti they provide additional constraints on the measurement of a galactic abundances of alpha-elements, i.e. elements produced by fusion of alpha-particles.


Maria Bergemann (MPA), Rolf Peter Kudritzki (IfA, Hawaii/MPA), Karin Lind (MPA)


References

Bergemann, Kudritzki, et al. 2012, ApJ, in prep.

Kudritzki, Urbaneja, Gazak et al. 2012, ApJ, 747, 15

Davies, Kudritzki, Figer 2010, MNRAS, 407, 1203

Gieren, Pietrzynski, Bresolin, Kudritzki et al. 2005, ESO Messenger, No.121, p. 23-28

Cushing, M.C., Rayner, J.T, & Vacca, W.D. 2005, ApJ, 623, 1115

Rayner, J.T., Cushing, M.C., & Vacca, W.D. 2009, ApJS, 185, 289