aa r X i v : . [ a s t r o - ph ] J un Hot Massive Stars: The Impact of HST
Paul A. Crowther
Dept of Physics & Astronomy, University of Sheffield, Hounsfield Road, Sheffield,S3 7RH, United Kingdom
Summary.
We review the contribution of Hubble Space Telescope to the studyof hot, luminous stars. Optical and IR imaging have permitted spatially resolvedobservations of young, massive clusters within Local Group galaxies, such as R136,NGC 3603 and Arches, revealing unprecedented concentrations of very massive Ostars. UV spectroscopy of field OB stars in the Magellanic Clouds have providedsuitable templates for interpretation of metal-poor star-forming galaxies at high-redshift. Spectacular imaging provides the detailed structure of ejecta nebulae fromindividual stars, including the Homunculus associated with η Carinae and M1–67,associated with a Wolf-Rayet star. HST has permitted individual massive stars tobe spatially resolved in giant HII regions located beyond the Local Group, such asNGC 604, plus individual clusters, dominated by the light of massive stars withinstarburst galaxies at larger distances, such as NGC 3125. UV spectroscopy of young,massive clusters in the extremely metal-poor HII galaxy I Zw 18 include signaturesof large numbers of Wolf-Rayet stars.
Massive stars distinguish themselves from their lower mass siblings by theirexceptionally high main-sequence luminosities, such that their lifetimes aremeasured in Myr rather than Gyr. Individual stars may be readily studied indetail in external galaxies - it was not be accident that the first post-servicingmission WFPC2 image was obtained of the Wolf-Rayet star Melnick 34 in the30 Doradus star-forming region of the LMC (News Release: STScI-1994-05).Here, selected HST results are discussed, together with their impact uponEuropean astronomy.
The first balloon and satellite missions provided the means of studying windsfrom hot stars, via ultraviolet P Cygni profiles, for which IUE provided acomprehensive sample of Milky Way OB stars (e.g. Howarth & Prinja 1989).Comparable quality spectroscopy of OB stars in the metal-poor MagellanicClouds required the superior throughput and spatial resolution of HST (Wal-born et al. 1995, 2000, Evans et al. 2004).
Crowther
Winds of early-type stars are predicted to be driven by radiation pressurethrough (CNO, Fe-peak) metal-lines (e.g. Vink et al. 2001), from which metal-poor stars are expected to possess lower density, slower winds. Indeed, windvelocities of early O stars in the SMC were established to be slower thanthose of comparable Galactic stars (e.g. Prinja & Crowther 1998), from CIV1550 observations, although it has taken large ground-based surveys, such asthe VLT/FLAMES survey of massive stars in Milky Way, LMC and SMCclusters to quantify the metallicity-dependence of mass-loss rates, revealingdM/dt ∝ Z . ± . (Mokiem et al. 2007), in good agreement with theory Z . ± . according to Vink et al. (2001). Amongst the many images obtained with HST, one of the most breathtakinghas been the WFPC2 images of Morse et al. (1998) of the Homunculus re-flection nebula, produced by the ‘eruption’ of the prototype Luminous BlueVariable (LBV) η Carinae during the mid-19th Century, and now illuminatedfrom within. The expanding emission lobes extend 8.5 ′′ (0.1 pc) from the cen-tral star, now known to be a 5.5 yr period binary system, for which WFPC2achieved a spatial resolution of ∼
115 AU, comparable with the size of ourSolar System. A physical mechanism for the eruption, in which in excess of10 M ⊙ were ejected over 20 years, remains unclear. STIS long-slit spectroscopyof the central star, reveals exceptional properties, with current mass-loss ratesof 10 − M ⊙ yr − for an adopted (infrared) luminosity of 5 × L ⊙ (Hillieret al. 2001).WFPC2 has also provided an unprecedented view of the young ejecta (orring) nebula M1–67 associated with the Galactic Wolf-Rayet star WR124(Grosdidier et al. 1998). The radial density distribution of this nebula, ap-proximately 90 ′′ ( ∼ r − . dependence, has enabledphoto-ionization modelling of the central WR star (Crowther et al. 1999) andmay represent the immediate environment into which cosmological Gamma-Ray Bursts (GRBs) explode. M1–67 is compared to η Carinae on a commonphysical scale in Figure 1.
High spatial resolution radio surveys of massive binaries reveal thermal (stel-lar wind) and non-thermal (colliding wind) components (e.g. Williams etal. 1997). Positions of stars within such systems have been established withWFPC2 imaging, enabling their relative wind strengths to be established(Niemela et al. 1998).Indeed, FGS has enabled searches for hitherto unknown massive binariesin regions of parameter space inaccessible to ground-based techniques. One ot Massive Stars 3
Fig. 1.
WFPC2 imaging, to scale, of the Homunculus nebula associated with η Carinae (left, 0.3 × × such survey of 23 OB stars in Carina revealed 5 new binaries, includingan apparent early O dwarf companion to HD 93129A, the prototype O2supergiant, separated by only 55 mas (137 AU, Nelan et al. 2004). Binaritywas later confirmed by detection of a non-thermal component in the observedradio emission.To date, the current heavyweight record holder is WR20a, a 3.7 day eclips-ing binary system composed of two H-rich WN-type stars each of ∼ M ⊙ (Rauw et al. 2005). Historically, R136 the central ionizing cluster of 30 Doradus in the LMC wasconsidered as a potential supermassive star. 30 Doradus is the brightest giantHII region within the Local Group, responsible for the equivalent of 1,000
Crowther equivalent O7 V stars. Weigelt & Baier (1985) first resolved the central sourceR136 into multiple components using speckle imaging. De Marchi et al. (1993)provided confirmation with FOC, while Massey & Hunter (1998) undertookFOS spectroscopy of individual sources within R136, revealing a multitude ofearly O stars, indicating extreme youth (1–2 Myr) and apparently very highindividual stellar masses, perhaps up to 120 M ⊙ . The total stellar mass ofR136 probably exceeds 5 × M ⊙ (Hunter et al. 1996). NICMOS imagingof 30 Doradus has identified still younger, visibly obscured, young massivestars. Fig. 2.
WFPC2 imaging, to scale, of the young star clusters NGC 3603 (4 × ×
10 pc, ESA heic0416, Hunteret al. 1996)
HST has also spatially resolved the Milky Way cluster NGC 3603, againrevealing many early O stars (Drissen et al. 1995). The central cluster iscomparable to R136a within a radius of ∼ < ∼ M ⊙ has been proposed (Figer 2005).At present, the most massive young Milky Way cluster known is Wester-lund 1 (Wd 1), in which ∼ × M ⊙ are contained within a radius of ∼ ot Massive Stars 5 Fig. 3.
HST imaging, to scale, of the giant HII regions NGC 346 (SMC N11A) inthe SMC (ACS: 90 ×
90 pc, STScI-2005-04, Nota et al. 2006) and NGC 604 in M 33(WFPC2: 360 ×
450 pc, STScI-1996-27: Yang et al. 1996) including Wolf-Rayet stars, red supergiants, yellow hypergiants, an LBV anda B[e] supergiant.Such high mass, compact clusters are unusual for normal spiral galax-ies. Large, scaled-up OB associations are more typical, for which WFPC2imaging provided an excellent example of NGC 604 in M33 (see Yang et al.1996), several hundred pc in diameter, containing the equivalent of severalhundred O7 V stars. This is compared to ACS imaging of NGC 346 in theSMC (Nota et al. 2006) in Fig. 3, somewhat smaller and with the equivalentof 50 equivalent O7 V stars (Kennicutt 1984). The high spatial resolutionachieved with WFPC2 permitted Drissen et al. (1993) to identify individ-ual Of and Wolf-Rayet emission line stars in NGC 604 to be established. Asimilar ground-based study of Wolf-Rayet stars within a bright giant HII re-gion in the southern barred spiral galaxy NGC 1313 by Hadfield & Crowther
Crowther (2007), comparable in size and ionizing output to NGC 604 demonstrates theinvaluable role of ACS in disentangling the stellar content. Field Wolf-Rayetstars at the 4 Mpc distance of NGC 1313 can be detected from ground-basedtelescopes, but HST has proved decisive in detection core-collapse supernovaprogenitors from pre-SN imaging (e.g. Smartt et al. 2004).
Fig. 4. × ′′ ACS imaging of the blue compact dwarf galaxy NGC 3125 (U, V,H α , GO ∼ M ⊙ ) clusters fromknots A and B, ∼
500 pc apart (Hadfield & Crowther 2006) are indicated. z galaxies Local starbursts, such as NGC 3125 (11 Mpc) host young massive clusters,significantly more massive than R136, as presented in Figure 4. Starburstsfrom such galaxies provide an order of magnitude higher ionizing outputthan NGC 604 from within a much smaller region, e.g. the equivalent of2,500 equivalent O7 V stars from knot A of NGC 3125. STIS ultraviolet spec-troscopy of individual clusters enables robust age estimates of such clus-ters, based on comparisons between the OB P Cygni features and spectralsynthesis predictions (e.g. Chandar et al. 2004), as shown in Figure 5 forNGC 3125-A1 from Hadfield & Crowther (2006). Host galaxies typically pos-sess Magellanic Cloud metallicities, such that the use of template LMC/SMCOB stars (recall Sect.2) has proved to be invaluable. ot Massive Stars 7 F l ux ( − e r g s − c m − Å − ) H e II S i I V C I V NGC3125−A1
Fig. 5.
Dereddened, slit-loss corrected STIS spectroscopy of NGC 3125–A1 (thinsolid line), together with a 2 × M ⊙ Starburst99 Magellanic Cloud 4 Myr instan-taneous burst synthetic spectrum (thick solid line), plus LMC WN5–6 templates atHe ii λ In extreme cases, far lower metallicities are sampled, such as the HIIgalaxy I Zw 18, possessing 1/30 Z ⊙ based upon the recently reduced Solaroxygen abundance. Brown et al. (2002) undertook STIS ultraviolet spec-troscopy across I Zw 18, revealing multiple clusters in which carbon-sequenceWolf-Rayet stars were observed. Such prominent WR signatures at low metal-licity led Crowther & Hadfield (2006) to conclude that I Zw 18 hosts substan-tial WR populations, wholly unexpected for single star evolutionary modelsat such low metallicity.The composite rest-frame UV spectra of z ∼ ∼ Perhaps the most significant unresolved aspect relating to high mass starsis their formation, for which future high-spatial resolution infrared and ra-dio may prove decisive. Violent mass ejections, such as that experienced by η Crowther
Carinae, producing the Homunculus nebula, also remains ill-explained. A fur-ther puzzle is how large numbers of Wolf-Rayet stars form in young, massiveclusters within very metal-poor galaxies such as I Zw 18.