Astron. Nachr. / AN 331, No.5, 459–473 (2010) / DOI 10.1002/asna.200911342 Dissecting galaxies with quantitative spectroscopy of the brightest stars in the Universe Karl Schwarzschild Award Lecture 2009 R.-P. Kudritzki⋆ Institute for Astronomy, University of Hawaii, 2680 Woodlawn Dr., Honolulu, Hawaii 96822, USA Received 2009 Dec 10, accepted 2010 Jan 15 Published online 2010 May 17 Key words galaxies: abundances – galaxies: distances – stars: abundances – stars: distances – stars: early-type Measuring distances to galaxies, determining their chemical composition, investigating the nature of their stellar popu- lations and the absorbing properties of their interstellar medium are fundamental activities in modern extragalactic as- tronomy helping to understand the evolution of galaxies and the expanding universe. The optically brightest stars in the universe, blue supergiants of spectral A and B, are unique tools for these purposes. With absolute visual magnitudes up to ∼ MV = −9.5 they are ideal to obtain accurate quantitative information about galaxies through the powerful modern meth- ods of quantitative stellar spectroscopy. The spectral analysis of individual blue supergiant targets provides invaluable information about chemical abundances and abundance gradients, which is more comprehensive than the one obtained from H II regions, as it includes additional atomic species, and which is also more accurate, since it avoids the systematic uncertainties inherent in the strong line studies usually applied to the H II regions of spiral galaxies beyond the Local Group. Simultaneously, the spectral analysis yields stellar parameters and interstellar extinction for each individual su- pergiant target, which provides an alternative very accurate way to determine extragalactic distances through a newly developed method, called the Flux-weighted Gravity–Luminosity Relationship (FGLR). With the present generation of 10 m-class telescopes these spectroscopic studies can reach out to distances of 10 Mpc. The new generation of 30 m-class telescopes will allow to extend this work out to 30 Mpc, a substantial volume of the local universe. c 2010 WILEY-VCH Verlag GmbH& Co. KGaA, Weinheim 1 Introduction systematic uncertainties, as we will show later in the course of this lecture. To measure distances to galaxies, to determine their chemi- Thus, is it really out of the question to apply the meth- cal composition, and to investigate the nature of their stellar ods of quantitative spectroscopyof individualstars as a most populations and the absorbing properties of their interstellar powerful complementary tool to understand the evolution of medium are fundamental activities in modern extragalactic galaxies beyond the Local Group? The answer is, no, it is astronomy. They are crucial to understand the evolution of not. In the era of 10m-class telescopes with most efficient galaxies and of the expanding universe and to constrain the spectrographs it is indeed possible to quantitatively analyze history of cosmic chemical enrichment, from the metal-free the spectra of individual stars in galaxies as distant as 10 universe to the present-day chemically diversified structure. Mpc and to obtain invaluable information about chemical However, while stars are the major constituents of galax- composition and composition gradients, interstellar extinc- ies, little of this activity is based on the quantitative spec- tion and extinction laws as well as accurate extragalactic troscopy of individual stars, a technique which over the last distances. With the even larger and more powerful next gen- fifty years has proven to be one of the most accurate di- eration of telescopes such as the TMT and the E-ELT we agnostic tools in modern astrophysics. Given the distances will be able to extend such studies out to distances as large to galaxies beyond the Local Group individual stars seem as 30 Mpc. All one has to do is to choose the right type of to be too faint for quantitative spectroscopy and, thus, as- stellar objects and to apply the extremely powerful tools of tronomers have settled to restrict themselves to the photo- NLTE spectral diagnostics, which have already been suc- metric investigation of resolved stellar populations, the pop- cessfully tested with high resolution, high signal-to-noise ulation synthesis spectroscopy of integrated stellar popula- spectra of similar objects in the Milky Way and the Magel- tions or the investigation of H II region-emission lines. Of lanic Clouds. course, color-magnitude diagrams and the study of nebu- Of course, now when studying objects beyond the Local lar emission lines have an impressive diagnostic power, but Group the analysis methods need to be modified towards they are also limited in many ways and subject to substantial medium resolution spectra with somewhat reduced signal. For a stellar spectroscopist, who is usually trained to believe ⋆ Corresponding author: [email protected] that only the highest resolution can give you an important c 2010 WILEY-VCH Verlag GmbH& Co. KGaA, Weinheim 460 R.P. Kudritzki: Dissecting galaxies answer, this requires some courage and boldness (or maybe oxygen emission lines using so-called strong-line methods, naive optimism). But as we will see in the course of this which, as we will show, have huge systematic uncertain- lecture, once one has done this step, a whole new universe ties arising from their calibrations. Direct stellar abundance is opening up in u the true sense of the word. studies of blue supergiants open a completely new and more accurate way to investigate the chemical evolution of galax- ies. 2 Choosing the right objects: A and B super- giants In addition, because of their enormous intrinsic bright- ness, blue supergiants are also ideal distance indicators. As first demonstrated by Kudritzki, Bresolin & Przybilla It has long been the dream of stellar astronomers to study (2003) there is a very simple and compelling way to use individual stellar objects in distant galaxies to obtain de- them for distance determinations. Massive stars with masses tailed spectroscopic information about the star formation in the range from 12 M to 40 M evolve through the history and chemodynamical evolution of galaxies and to ⊙ ⊙ B and A supergiant stage at roughly constant luminosity. determine accurate distances based on the determination of In addition, since the evolutionary timescale is very short stellar parameters and interstellar reddening and extinction. when crossing through the B and A supergiant domain, the At the first glance, one might think that the most massive amount of mass lost in this stage is small. As a conse- and, therefore, most luminous stars with masses higher than quence, the evolution proceeds at constant mass and con- 50 M are ideal for this purpose. However, because of their ⊙ stant luminosity. This has a very simple, but very important very strong stellar winds and mass-loss these objects keep consequence for the relationship between gravity and effec- very hot atmospheric temperatures throughout their life and, tive temperature along each evolutionary track, namely that thus, waste most of their precious photons in the extreme 4 the flux-weighted gravity, gF = g/T , stays constant. As ultraviolet. As we all know, most of these UV photons are eff shown in detail by Kudritzki et al. (2008) and explained killed by dust absorption in the star forming regions, where again further below this immediately leads to the ‘flux- these stars are born, and the few which make it to the earth weighted gravity–luminosity relationship’ (FGLR), which can only be observed with tiny UV telescopes in space such has most recently proven to be an extremely powerful to as the HST or FUSE and are not accessible to the giant tele- determine extragalactic distances with an accuracy rivalling scopes on the ground. the Cepheid- and the TRGB-method. Thus, one learns quickly that the most promising objects for such studies are massive stars in a mass range between 15 to 40 M⊙ in the short-lived evolutionary phase, when 3 Spectral diagnostics and studies in the they leave the hydrogen main-sequence and cross the HRD in a few thousand to ten thousand years as blue supergiants Milky Way and Local Group of B and early A spectral type. Because of the strongly re- duced absolute value of bolometric correction when evolv- The quantitative analysis of the spectra of these objects is ing towards smaller temperature these objects increase their not trivial. NLTE effects and but also the influence of stel- brightness in visual light and become the optically brightest lar winds and atmospheric geometrical extension are of cru- “normal” stars in the universe with absolute visual magni- cial importance. However, over the past decades with an tudes up to MV =∼ −9.5 rivaling with the integrated light effort of hundreds of person-years sophisticated model at- brightness of globular clusters and dwarf spheroidal galax- mosphere codes for massive stars have been developed in- ies. These are the ideal stellar objects to obtain accurate cluding the hydrodynamicsof stellar winds and the accurate quantitative information about galaxies. NLTE opacities of millions of spectral lines. Detailed tests The optical spectra of B- and A-type supergiantsare rich have been carried out reproducing the observed spectra of in metal absorption lines from several elements (C, N, O, Milky Way stars from the UV to the IR and constraining Mg, Al, S, Si, Ti, Fe, among others). As young objects they stellar parameters with unprecedentedaccuracy; see reviews represent probes of the current composition of the interstel- by Kudritzki(1998),Kudritzki & Puls (2000), and Kudritzki lar medium. Abundance determinations of these objects can & Urbaneja (2007). For instance, the most recent work on therefore be used to trace the present abundance patterns in A-type supergiants by Przybilla et al. (2006, 2008), Schiller galaxies, with the ultimate goal of recovering their chem- & Przybilla (2008),and also by Przybilla et al. (2000),Przy- ical and dynamical evolution history.