The Galactic O-Star Spectroscopic Survey (GOSSS)
J. Maíz Apellániz, A. Sota, N. R. Walborn, E. J. Alfaro, R. H. Barbá, N. I. Morrell, R. C. Gamen, J. I. Arias, M. Penadés Ordaz
Highlights of Spanish Astrophysics VI, Proceedings of the IX Scientific Meeting of the Spanish Astronomical Society held on September 13 - 17, 2010, in Madrid, Spain. M. R. Zapatero Osorio et al. (eds.)
The Galactic O-Star Spectroscopic Survey (GOSSS)
J. Ma´ız Apell´aniz , A. Sota , N. R. Walborn , E. J. Alfaro , R. H. Barb´a , N. I.Morrell , R. C. Gamen , and J. I. Arias , Instituto de Astrof´ısica de Andaluc´ıa-CSIC, Granada, Spain Space Telescope Science Institute, Baltimore, MD, USA Departamento de F´ısica, Universidad de La Serena, La Serena, Chile Las Campanas Observatory, La Serena, Chile Instituto de Astrof´ısica de La Plata (CONICET, UNLP), La Plata, Argentina
Abstract
We present a massive spectroscopic survey of Galactic O stars, GOSSS, based on new,high signal-to-noise ratio, R ∼ O-type stars play a crucial role in the dynamic and chemical evolution of galaxies. Theyare the major source of ionizing and UV radiation and, through their huge mass-loss rates,they have a strong mechanical impact on their surroundings. Massive stars are also importantbecause the nuclear reactions in their interiors create a large fraction of the heavier chemicalelements. These nuclear products are blasted out into space in the final supernova explosionsthat put an end to the massive stars’ lives. Despite their importance, our knowledge of theseobjects and of their evolution is still incomplete because of their relatively small numbers,average large extinctions caused by their concentration on the Galactic plane, and the manyhidden or poorly studied multiple systems (Mason et al., 1998).In 2004 we compiled the most complete Galactic O-star Catalog (GOSC) with accu-rate spectral types ever assembled (Ma´ız Apell´aniz et al., 2004). GOSC was subsequently a r X i v : . [ a s t r o - ph . S R ] O c t The Galactic O-Star Spectroscopic Survey expanded from ∼
400 objects to the current ∼ To remedy the incompleteness of our knowledge of the basic properties of massivestars, in 2007 we started the Galactic O-Star Spectroscopic Survey (GOSSS). Its immediateobjective is to create the largest possible blue-violet (3900-5100 ˚A) spectroscopic database forvisually observable (
B <
14) Galactic O stars using high-S/N ( ≥ R ∼ ∼
700 of those 2500 objects using three different telescopes:the 1.5 m at Observatorio de Sierra Nevada (OSN), the 2.5 m duPont telescope at Las Cam-panas Observatory, and the 3.5 m telescope at Calar Alto (CAHA). The data are processedusing a devoted pipeline (Sota & Ma´ız Apell´aniz, these proceedings). Our goal is to observeat least one epoch for ∼ B <
8. Our goal isto observe all Galactic O stars down to B = 10 and possibly deeper.GOSSS is an ambitious project in terms of sample size and data quality but not inwavelength coverage, spectral resolution, or number of observed epochs, where other sur-veys provide better results. For example, its spectroscopic-binary detection capabilities andusefulness for detailed atmospheric modeling are limited. For those reasons, several of theauthors here are also involved in two other surveys, OWN (Barb´a et al., 2010) and IACOB(Sim´on-D´ıaz et al., 2010) that are observing a subset of the GOSSS stars at higher spectralresolution and multiple epochs. OWN covers southern stars and IACOB northern stars. Also,another survey whose PI is the same one as that of GOSSS is using Lucky Imaging to obtainhigh-resolution images of another subset of GOSSS stars (Ma´ız Apell´aniz, 2010). Taken alltogether, the four projects will provide an unprecedented window into the Galactic O starsin the solar neighborhood. a´ız Apell´aniz et al. The first GOSSS scientific results were published in Walborn et al. (2010) and Sota etal. (2010, ApJS submitted). Here are some of our findings: • We discovered C iii λλ ∼ • The peculiar category of Of?p stars was introduced by Walborn (1972) and for almost40 years it included just three Galactic members: HD 108, HD 148 937, and HD 191 612.The interest in the class has been recently revived by the discovery of magnetic fields inall of them (Donati et al., 2006; Martins et al., 2010; Wade et al., 2010). In the GOSSSdata we have found two new Of?p stars, NGC 1624-2 and CPD -28 o • We have presented a high-quality O-star atlas that can be used for spectral classification.The atlas has higher S/N and fewer gaps than previous examples and is accompaniedby MGB, a software that can be used to compare observed data with the standards,thus facilitating classification. • We have adopted a uniform horizontal classification criterion for late O stars based onthe He ii λ i λ ii λ i λ υ Ori, the former B0 V standard. • With respect to the vertical classification criteria, the new data have allowed us todistinguish luminosity class IV for spectral classes O6–O8 for the first time and todiscover several new examples of type O Vz. • Other changes to the classification criteria include the elimination of the ((n)) qualifier(line width, existing cases have become either (n) or normal) and of the f+ qualifier(presence of Si iv λ The Galactic O-Star Spectroscopic Survey • We have discovered some interesting correlations between extinction, DIB intensity andthe equivalent width of the interstellar Ca ii λ The final objectives of GOSSS run along five different research lines:
1. The fundamental spectral morphology of Galactic O stars.
The previously publishedlibraries of digital spectrograms of O stars at R ∼ <
100 O stars in total.Furthermore, those libraries include only a few northern stars and the quality of their datacan be significantly improved with modern CCD detectors. An improvement in size (from100 to 2000 O stars), quality (from S/N of 50 to 300), and uniformity of a spectral librarysimilar to the one in this project has always led in the past to new discoveries in the observedspectral morphologies. Those discoveries have produced, first, changes in the classificationcriteria and, later on, significant advances in the understanding of the underlying physicalprinciples through comparison with atmosphere models. We expect those two steps to takeplace in this project as well and some examples have been presented in the previous section.
2. The multiplicity of massive stars.
Most, if not all, massive stars are born in multiplesystems (Mason et al., 1998). This represents both a blessing and a curse for their study.It is a blessing because, at least in principle, it allows the measurement of their masses.The curse comes from the difficulty in doing so: long-period (thousands of years or more)Galactic massive binaries usually require high-resolution imaging or interferometry whileshort-period (days to months) ones require time-consuming multiple-epoch spectroscopy. In-between objects, those with unfavorable orientations, and high-order multiples with complexhierarchical orbits remain undetected (hence, with incorrect estimated masses) or with onlyuncertain values, thus hampering the derivation of the IMF. GOSSS, in conjunction with theOWN, IACOB, and our high-resolution imaging survey will attack this problem in a threefoldway: [1] observing with multiple-epoch spectroscopy at high and intermediate resolution (withthis approach we are following the orbits of several tens of massive multiple systems, manyof them previously unknown); [2] observing with high-resolution imaging; and [3] compilinginformation from the literature. This should yield the most complete ever study of themultiplicity of Galactic massive stars. Our preliminary OWN results are spectacular: out of240 O and WN stars we have detected radial velocity variations above 10 km/s for more than100 of them and for 26 objects in the sample we already have calculated spectroscopic binaryorbits (Barb´a et al., 2010). This raises the fraction of spectroscopic multiples among massivestars from ∼
20% to ∼
3. The optical-IR extinction law and the nature of the DIBs.
Nowadays, the most usedinterstellar extinction laws are those from Cardelli et al. (1989). The situation with theirquality is similar to that of O-star spectral libraries mentioned in the first research line: they a´ız Apell´aniz et al. E ( B − V ) < .
3) that can be greatly improved with modern means, thus overcoming some of the knownissues with the Cardelli laws. Here, we propose doing just that, combining the optical-NIRphotometry for the stars in GOSC with our newly derived spectral types using CHORIZOS(Ma´ız Apell´aniz, 2004) and the newly recalibrated photometric zero points (Ma´ız Apell´aniz,2005, 2006, 2007). The newly derived extinction laws will have variable NIR slopes. Thecombination of a large library of high S/N optical O-star spectra (from GOSSS, OWN, andIACOB) with an improved knowledge of the amount and type of extinction for each star willalso allows us to analyze the decades-old problem of the nature of the DIBs (Herbig, 1995)with a fresh impulse.
4. The spatial distribution of massive stars and dust within 3 kpc of the Sun.
Mostrecent studies of the spatial distribution of Galactic extinction have concentrated on either[a] the diffuse component at high latitudes and low extinction (e.g. Juri´c et al. 2008) or[b] the high extinction regions located behind molecular clouds or at large distances (e.g.Alves et al. 2001) because those studies can be effectively carried out using large photometricsurveys where the target stars are of A and later type. On the other hand, the once popularsurveys based on OB-star spectroscopy (e.g. Neckel et al. 1980) have faded out of style,leaving the region within 3 kpc of the Sun mostly unstudied even though the quality of theavailable data is much better than 30 years ago. Our data, with the aid of completenessinformation from 2MASS and (in the future) WISE, should cover this important gap andprovide the best available information on the spatial distribution of massive stars and dustwithin 3 kpc of the Sun.
5. The massive end of the IMF in the solar neighborhood.
Once the four previousresearch lines have reached their major objectives (accurate spectral types, characterizationof multiple systems, and precise extinction corrections and spectroscopic distances for oursample) we will attack one of the current holy grails of Galactic research, the IMF for massivestars, with data of unprecedented quality and overcoming many of the biases present inprevious attempts. Our goal here is to redo the work of Massey et al. (1995) with betterdata, a larger sample, and improved numerical techniques.
Acknowledgments
Support for this work was provided by [a] the Spanish Ministerio de Ciencia e Innovaci´on throughgrants AYA2007-64052 and AYA2010-17631, the Ram´on y Cajal Fellowship program, and FEDERfunds; [b] the Junta de Andaluc´ıa grant P08-TIC-4075; and [c] NASA through grants GO-10205, GO-10602, and GO-10898 from STScI, which is operated by AURA Inc., under NASA contract NAS 5-26555. This research has made extensive use of Aladin (Bonnarel et al., 2000); the SIMBAD database,operated at CDS, Strasbourg, France; and the WDS Catalog maintained at USNO (Mason et al., 2001).
References
Alves, J. F., Lada, C. J., & Lada, E. A. 2001, Nature, 409, 159