EDP Sciences
Free Access
Issue
A&A
Volume 584, December 2015
Article Number A5
Number of page(s) 19
Section Galactic structure, stellar clusters and populations
DOI https://doi.org/10.1051/0004-6361/201525882
Published online 13 November 2015

© ESO, 2015

1. Introduction

Our knowledge of the massive-star populations of the Magellanic Clouds has increased significantly over the past decade, largely via observations with multi-object spectrographs (e.g. Massey & Olsen 2003; Evans et al. 2004, 2006, 2011; Fariña et al. 2009; Lamb et al. 2013). Such surveys have been used to address questions pertaining to stellar evolution (e.g. Massey & Olsen 2003; Evans et al. 2008), wide-area studies of stellar kinematics (e.g. Evans & Howarth 2008), the formation of massive stars in relative isolation (e.g. Lamb et al. 2010; Bressert et al. 2012; Oey et al. 2013), the structure of stellar clusters (e.g. Hénault-Brunet et al. 2012a,b), the properties of the interstellar medium in the Clouds (e.g. Welty et al. 2006; van Loon et al. 2013), and, via multi-epoch observations, the binary properties of massive stars (e.g. Bosch et al. 2009; Sana et al. 2013; Dunstall et al. 2015).

thumbnail Fig. 1

Location of our AAOmega targets overlaid on a blue-optical image from the Digitized Sky Survey (DSS). The ten early-type stars (with classifications of O4 or earlier) discussed in Sect. 3 are highlighted in red.

Open with DEXTER

The compendia of known spectral types of massive stars in the Clouds by Bonanos et al. (2009, 2010) emphasised the disparity in our knowledge of the spectral content of the Clouds, with classifications for 5324 stars in the Small Magellanic Cloud (SMC), but only 1750 in the Large Magellanic Cloud (LMC). Filtering the LMC catalogue from the Magellanic Clouds Photometric Survey (MCPS, Zaritsky et al. 2004) by a faint magnitude limit of V = 15.3 mag and a colour cut of (VI)< 0.0 mag to identify likely O- and luminous B-type objects gives a total of ~10 000 stars1. The true number of early-type stars in the LMC will be influenced by the effects of crowding in the MCPS photometry and the number of interlopers included via the above filters, but there clearly remains much to learn about the massive-star content of the LMC, even in light of recent classifications for 780 O- and B-type stars in 30 Doradus (Walborn et al. 2014; Evans et al. 2015).

In this article we present spectroscopy from observations in early 2006 (P.I. van Loon) to test the capabilities of the then new AAOmega spectrograph (Saunders et al. 2004; Sharp et al. 2006) on the 3.9 m Anglo-Australian Telescope (AAT). Two fields were observed in the LMC, one in the NE near the massive 30 Doradus star-forming region and a second in the NW, centred on the N11 region. Spectra for thirteen of the targets have been published to date: seven peculiar “nfp” stars (Walborn et al. 2010), a massive runaway O2-type star (Evans et al. 2010), and five eclipsing binary systems (Muraveva et al. 2014). Partly motivated by results from the VLT-FLAMES Tarantula Survey (VFTS, Evans et al. 2011) regarding runaway stars (Evans et al. 2010, 2015), we have revisited the AAOmega data in the NE field to investigate the spectral content and dynamics of massive stars in this part of the LMC.

In this paper we present spectral classifications for 263 stars in the north-eastern region of the LMC, only 60 (23%) of which have previous classifications (and often from observations of much lower quality and/or resolution). We describe the characteristics and reductions of the observations, including the optical photometry of our sources, in Sect. 2. The spectral classifications of our targets are discussed in Sect. 3, and estimates of stellar radial velocities (RVs) are presented in Sect. 4, followed by a short summary in Sect. 5. The observations in the field centred on the N11 region will be presented in a future article.

2. Observations

AAOmega is a fibre-fed, twin-arm spectrograph, with light separated into blue and red arms using a dichroic beam splitter (Saunders et al. 2004; Sharp et al. 2006). Up to 392 objects across a field on the sky that is 2° in diameter can be observed simultaneously, using fibres on the prime-focus focal plate configured with the robotic positioner of the Two-degree Field facility (2dF, Lewis et al. 2002); each fibre has an on-sky aperture of 2′′. Targets are selected in the 2° field using the configure software (see Lewis et al. 2002).

Luminous early-type stars were selected as potential targets from the MCPS (Zaritsky et al. 2004) using a magnitude cut of V< 14 mag (to ensure a signal-to-noise ratio of >50) and a colour cut of (VI) < 0.0 mag to identify early-type stars. We used configure to select targets from our input list using a field centre of α= 05h36m07s, δ=69°1558′′ (J2000). For context, these coordinates are 3 NE of the Honeycomb nebula (Wang 1992), 3.́5 E of SN1987A, and 17 SW of R136 (the massive cluster at the centre of 30 Dor). The large field (2° diameter) enabled us to observe a reasonable number of targets in the regions immediately south of 30 Dor (including NGC 2060), in the N154, N158, and N160 complexes further to the south (Henize 1956), and field stars across the region. To illustrate the distribution of our targets, their locations are overlaid on the Digitized Sky Survey (blue-optical) image in Fig. 1.

The NE LMC field was observed on 2006 February 22–23. The data presented here were obtained with the blue arm, using the 1700B grating on the first night, and the 1500V grating on the second. Simultaneous observations were obtained with the red arm using the 1700D grating and centred at 8620 Å; this region contains fewer lines of interest for massive stars than the blue spectra so these data were not considered further.

The data were reduced using the 2dfdr software (Lewis et al. 2002). In brief, 2dfdr was used for bias subtraction, fibre location, extractions, division by a normalised flat-field, and wavelength calibration of each target. Subsequent processing included correction of the spectra to the heliocentric frame, sky subtraction, rejection of significant cosmic rays, and preliminary normalisation (using pre-defined continuum regions). The delivered spectral coverage and resolution from the observations with the blue arm of the spectrograph is summarised in Table 1. The signal-to-noise ratio of the final spectra obtained with the 1700B grating is 50–60 per rebinned pixel for the faintest targets, and in excess of 100 for the brightest supergiants. The signal-to-noise ratio of the (longer and slightly lower resolution) 1500V observations is typically 20–30 greater than for the 1700B data.

Astrometry and optical photometry for each target (from the MCPS) is listed in Tables 2 and 3, respectively – the identifiers in the first column are simply the running numbers (in ascending RA) from our list of potential targets (hence those observed do not run in a continuous sequence); for consistency with the identifications used by Walborn et al. (2010) we also adopt them here. We note that the majority of the photometry from Zaritsky et al. for stars with V< 13.5 mag was taken from the survey by Massey (2002), which employed a photometric aperture of 16.̋2; thus in many instances crowding and/or nebular contamination may well influence the values in Table 3 (e.g. see discussion by Evans et al. 2011). Cross-matches of our targets with identifications/aliases from past spectroscopy are included in the final column of Table 2.

Table 1

AAOmega spectrograph settings used.

3. Spectral classification

The AAOmega spectra were classified by visual inspection in comparison with standards following the usual precepts for early-type stars (Walborn & Fitzpatrick 1990; Sota et al. 2011, 2014), taking into account the reduced metallicity of the LMC (e.g. Fitzpatrick 1988; Walborn et al. 1995, 2014; Evans et al. 2015), the effects of rotational broadening and the spectral resolution of our data. In brief, the primary diagnostic lines in the O-type spectra are the ionisation ratios of the helium lines, while also taking into account absorption from Si III at the latest types. The additional qualifiers employed in the classifications are summarised in Table 3 of Sota et al. (2011), and the spectral-type and luminosity-class criteria used for the later types are those summarised in Tables 4–6 of the same study. The B-type classifications employed the same spectral-type criteria as those in Tables 1 and 2 of Evans et al. (2015), with luminosity classes assigned from the width of the Balmer lines, while also taking into account the intensity of the silicon absorption lines at the earlier types; example sequences for O- and B-type spectra were given by Sota et al. (2011) and Evans et al. (2015), respectively.

The classifications of the AAOmega spectra are listed in Table 2, representing the first classifications for 203 of our targets. Previous classifications for the remaining 60 stars are summarised in Table 4 (complete to the best of our knowledge); in many cases the AAOmega spectra are superior to past spectroscopy, e.g. the resolution of the spectra obtained by Testor & Niemela (1998) was only 8 Å. Initial inspection of the spectra revealed eleven double-lined binaries (SB2s), and a number of candidate single-lined binaries (SB1s), discussed further in Sect. 4.1.

Our classifications of the B-type supergiants include suffixes to indicate nitrogen lines which are strong (Nstr) or weak (Nwk) compared to morphologically-normal stars of the same adopted spectral type. These qualifiers were diagnosed from visual inspection of the CNO absorption features throughout the spectra, but principally informed by the intensity of the N II λ3995 and the CNO features in the λλ4640–4650 region (Walborn 1976; Fitzpatrick 1991). An example of the contrast between Nstr and Nwk spectra of B-type supergiants is given in Fig. 1 of Evans et al. (2015).

The sample includes ten stars with classifications of O4 or earlier and their spectra are shown in Fig. 2. Four of these were previously unknown, namely: AAΩ 30 Dor 101, 181, 248, and 280. Given the absence of He I λ4471 in the spectrum of AAΩ 30 Dor 159, we adopt a classification of O3.5 III(f) over the O4 III(f) from Walborn et al. (2002a). The sample also includes examples of the nfp class of peculiar O-type spectra, which are defined by composite emission and absorption in the He II λ4686 line (Walborn 1973). The six nfp stars in the sample, AAΩ 30 Dor 142, 187, 320, 333, 368 and 380, were included in the discussion of the phenomenon by Walborn et al. (2010), and we adopt their classifications here.

thumbnail Fig. 2

AAOmega spectra of the ten early O-type stars in the sample (with each spectrum smoothed by a 5-pixel median filter for clarity and offset by 0.5 continuum units). Absorption lines identified in the spectrum of AAΩ 30 Dor 364 are: He II λλ4026, 4200, 4542, 4686; He I λ4471; N V λλ4604, 4620. The emission lines identified in the spectra are, in order of increasing wavelength: N IV λ4058; Si IV λλ4089, 4116; N III λλ4634-40-42; C IV λ4658. Broad absorption from the λ4430 diffuse interstellar band can be seen in some spectra (e.g. AAΩ 30 Dor 254).

Open with DEXTER

The classification of AAΩ 30 Dor 078 (SOI 399, Stock et al. 1976) also merits brief discussion. Classified as A0 Ia by Stock et al., its classification here as B3 Iab is tantalysing in the context of variations associated with luminous blue variables. The objective-prism spectroscopy used by Stock et al. was relatively coarse in terms of spectral resolution, but it is notable that their classifications include spectra as early as B6 (for comparable magnitudes and from the same photographic plate), so it is plausible to assume that they would have been able to distinguish SOI 399 as a B-type spectrum if it was in the same state as the 2006 observations2. Checks of the ASAS-3 database (spanning 2000–2009, see Pojmański 2002) and the DASCH archive (spanning 1890–1990, Grindlay et al. 2012) reveal no significant photometric variations (given the cadence of the available data).

3.1. Indications of CNO abundances in early O-type stars

From inspection of the spectra in Fig. 2 we noted the apparent weakness of the nitrogen features in AAΩ 30 Dor 248 and 280. These lines are the primary classification criteria at such early types (see Walborn et al. 2002a), yet in these two spectra there is only weak/marginal N III λλ4634-40-42 and N IV λ4058 emission, and an absence of N V λλ4604-4620 absorption. The absence/weakness of He I λ4471 (and other He I features) argues for the early-types adopted here, and we suggest these are the nitrogen-poor counterparts of the morphologically normal and nitrogen-rich O2-type stars discussed by Walborn et al. (2004).

Nitrogen enrichment/deficiency in the spectra of late O- and early B-type spectra was first noted by Walborn (1976), primarily with reference to the absorption strengths of the CNO features in the λλ4640–4650 region. At earlier types, from both morphlogical considerations and quantitative analysis, Walborn et al. (2004) argued that some O2-type spectra (classified as ON2) are nitrogen-rich compared to morphologically normal O2-type spectra, suggesting a more advanced evolutionary state or greater chemical enrichment via initially-larger rotational velocities.

By analogy to the ON/OC sequence at later types (e.g. Walborn 1976; Sota et al. 2011), we therefore classify AAΩ 30 Dor 248 and 280 as OC-type spectra. As well as the chemical abundances, a broad range of physical factors (e.g. gravity and mass-loss rate) influence the appearance of the nitrogen features in the earliest O-type stars (Rivero González et al. 2012). On morphological grounds we employ the OC classification for the two spectra with an absence of N V λλ4604, 4620 absorption, combined with C IV λ4658 emission. In addition, the weak but still discernable N V absorption in the spectrum of AAΩ 30 Dor 333 argues for an Nwk qualifier (Walborn, priv. comm.). As at later types, we suggest that AAΩ 30 Dor 248 and 280 could be examples of an earlier evolutionary stage (i.e. less chemical processing/enrichment) than the morphologically-normal objects such as AAΩ 30 Dor 254, with the ON2 star, AAΩ 30 Dor 040, completing the sequence. To highlight the trend in the N V absorption lines in this sequence (and the presence of C IV emission), a subset region of the AAOmega spectra for three stars is shown in Fig. 3.

thumbnail Fig. 3

λλ4500-4725 range of three early-type AAOmega spectra, illustrating the large variations in N V λλ4604-20 absorption and the presence of C IV λ4658 (and O IV λ4632) emission.

Open with DEXTER

Pending further observations (including the Hα profiles to constrain the wind properties), we calculated investigative synthetic spectra using the PoWR model atmosphere code (Hamann 2003, 2004). Initially developed for analysis of Wolf–Rayet type spectra, the code is also well suited for analysis of O-type stars (e.g. Oskinova et al. 2007; Evans et al. 2012). Adopting the physical properties from the published analysis of VFTS 016 (i.e. AAΩ 30 Dor 254) from Evans et al. (2010), we calculated seven PoWR models to investigate if the C IV λ4658 line is genuinely sensitive to abundance, or if it is significantly affected by other physical parameters.

The baseline model parameters were: an effective temperature (Teff) of 50 kK, luminosity (L) of log (L/L) = 6.08, gravity (ggrav) of log ggrav = 3.75, microturbulence (ξ) of 30 km s-1, and a stellar wind with a terminal velocity (v) of 3450 km s-1, an acceleration law described by a β parameter of 1.0, and with a mass-loss rate () of 10-5.5M yr-1. Chemical abundances (by mass fraction) were XH = 0.7374, XHe= 0.258, XN= 0.0008, XC = 0.0008, XO= 0.0016. In the six panels of Fig. 4 we show the effects, from top to bottom, of reducing the mass-loss rate to 10-7M yr-1, reducing the terminal velocity to 1000 km s-1, introducing a clumped wind with a volume-filling factor, f= 0.1 (i.e. 10%), decreasing the microturblence to 15 km s-1, reducing the iron abundance by a factor of two, and reducing the carbon abundance by a factor of ten; changing β by ±0.2 also had minimal impact. All of the spectra have been convolved with a rotational broadening profile of 150 km s-1, to match that adopted by Evans et al. (2010). These preliminary tests suggest that the carbon abundance is the principal factor influencing the appearance of the C IV line (for this adopted temperature and luminosity), and that our hypothesis of these stars as carbon-rich (relative to nitrogen) is plausible3.

thumbnail Fig. 4

Synthetic PoWR spectra (black line) adopting parameters for AAΩ 30 Dor 254 from Evans et al. (2010). The spectra plotted in red, moving from the top to bottom, are models which vary the mass-loss rate (Mdot), terminal velocity (v) clumping factor (fcl), microturbulence (ξ), iron abundance (XFe, to investigate possible blanketing effects), and carbon abundance (XC), as detailed in Sect. 3.1. The lines identified in the lower panel are He II λλ4542, 4686; N V λλ4604-20, and C IV λ4658.

Open with DEXTER

4. Stellar radial velocities

To estimate RVs for the O- and B-type stars we used Gaussian fits of the absorption lines listed in Table 5. The lines selected are those used by Sana et al. (2013) and Evans et al. (2015) to analyse the VFTS data, although we chose not to use He λ4026 (blend of He I and II) nor He I λ4121 (blend with O II). We note that He I λ4471 was also avoided as it is typically the line most affected by nebular contamination, and also presents other discrepancies for RV estimates, probably due to its nature as a triplet transition and a possible blend with O II (for further discussion see Appendix B of Sana et al. 2013).

These adopted lines were not available at the earliest and latest types in the AAOmega sample. Thus, for the earliest O-type stars (O4) we used He II λλ4026, 4200, 45424. For the later-type (B9 and A-type) spectra we used Hδ and Hγ, combined with Si II λλ4128-31 and He I λ4471.

Mean RVs for each target for the two nights, v1 and v2, are presented in Table 2. The quoted uncertainties are the standard errors (s.e.) on the means, and the sixth and eighth columns give the number of measured lines (n); mean values are only calculated for spectra where n ≥ 3. As a check for systematics in our measurements (e.g. line blends) and the adopted rest wavelengths, for each line in our analysis we calculated the mean residuals (ΔNight1 and ΔNight2, and their standard deviations) compared to the estimated RV for each star; these are given in the third and fourth columns of Table 5. In general, there are no significant systematic offsets present in the adopted lines. The Balmer lines in the cooler spectra appear to yield (marginally) different RV estimates, but these values are sufficient for our purposes.

Table 5

Rest wavelengths used for the radial-velocity estimates.

4.1. Binaries

Our observations were obtained on two (consecutive) nights so our velocity estimates are somewhat limited in the search for variations arising from single-lined binaries. Nonetheless, we employed similar criteria to those used by Sana et al. (2013) and Dunstall et al. (2015) for identification of spectroscopic binaries from the VFTS. For stars with RV estimates from both nights, we consider it as a spectroscopic binary if they satisfy that | v1v2 |> 20 km s-1 between the observations and that | v1v2 |/, i.e. such variations are statistically significant, following the approach taken by Sana et al. (2013). The choice of the RV threshold for binary detection is a trade-off between the number of false positives arising from the effects of pulsations (and other atmospheric variations) and detections of real RV shifts from binary motion, as discussed by Sana et al. (2013) and Dunstall et al. (2015).

Employing these criteria, the RV estimates for only one star, AAΩ 30 Dor 053 (classified as B0.5 Ib Nwk), is formally significant as a SB1 system; this remains the case using a lower threshold of | v1v2 | > 16km s-1 (as used by Dunstall et al. 2015, in their analysis of early B-type stars). There are a further six stars5 with | v1v2 | > 20km s-1, but with sufficient uncertainties on their RV estimates that they do not satisfy the second criterion; for the purposes of the calculations in the next section we exclude these potential (though unconfirmed) RV variables.

In summary, eleven SB2 systems were found in our spectroscopy (with an additional SB2 candidate, AAΩ 30 Dor 173), one SB1 system (AAΩ 30 Dor 053), and six potential RV variables; these objects and their spectral classifications are listed in Table 6. Eight are known eclipsing systems from Graczyk et al. (2011) from the third phase of the Optical Gravitational Lensing Experiment (OGLE). Using a 2′′ search radius we then cross-matched our remaining targets with the Graczyk et al. (2011) catalogue, finding eight other eclipsing systems that were undetected as binaries from the available spectra. The OGLE identifiers, periods, and light-curve classifications are included in Table 6. Similar cross-checks with the luminous variables in the LMC reported by Szczygieł et al. (2010) yielded only one match – AAΩ 30 Dor 286 (Sk69°238), classified as O7.5 Ib(f), which was detected as a low-amplitude (~0.25 mag) photometric variable.

Table 6

Summary of known binaries and candidate radial-velocity variables in the AAOmega sample.

4.2. Velocity distributions and radial-velocity outliers

We calculated mean velocities of all our targets with RV estimates, and for subsamples limited to the O- and BA-type spectra. The resulting means and their associated standard deviations are summarised in Table 7, and the systemic value is in good agreement with results from the VFTS (Evans et al. 2015). In calculation of these results we have excluded stars flagged as potential RV variables in the previous section and those identified as eclipsing binaries. We also identified and (by iteration) excluded RV outliers for both nights, defined as stars with: .

Table 7

Mean radial velocities and dispersions for the AAΩ sample for both nights ( and , respectively), excluding known and candidate binaries (see Sect. 4.1) and stars with outlying velocities (Table 8).

The RV estimates of the nine outliers are summarised in Table 8. Both AAΩ 30 Dor 254 and 383 only have estimates (from three or more lines) for one night, but qualitative comparison of the spectra from the two nights reveals no obvious shifts; indeed, the former of these two objects is VFTS 016, identified as a runaway star by Evans et al. (2010, including analysis of these data).

For completeness, we note that two stars, AAΩ 30 Dor 248 and 282, are (marginal) outliers in the estimates from one of the nights, suggesting either small RV variations or simply that we are at the limit of the available data (given that relatively few lines were available for RV estimates for these stars). Following the suggestion by Walborn et al. (2002a) that Sk68°137 and BI 253 (VFTS 072, AAΩ 30 Dor 276) might be massive runaways, the case of AAΩ 30 Dor 248 is similarly intriguing given its location (~22 NNW of R136, see Fig. 1). Indeed, if its OC-type classification does indicate an early evolutionary phase it raises the question of whether it is an ejected runaway, or formed more locally (in relative isolation given the apparent lack of a nearby star-forming region).

AAΩ 30 Dor 159 (W61 28-23) is the second largest outlier, and we note that Massey et al. (2005) reported a comparable RV (~350 km s-1, see their Fig. 9) from observations in 1999 January. Whether this is a genuine runaway, or just chance observations of similar RVs for a large-amplitude binary system will require further spectroscopic monitoring. Indeed, additional spectroscopy of each object in Table 8 will be required to ascertain their true status.

Table 8

Stars with outlying radial velocities, which are candidate large-amplitude binaries or runaway stars.

5. Summary

We have presented spectral classifications from optical spectroscopy with 2dF-AAOmega for 263 massive stars in the NE region of the LMC, together with RV estimates for 233 stars in the sample. Ten stars have classifications of O4 or earlier, with two (AAΩ 30 Dor 248 and 280) classified as OC-type given the nitrogen deficiency of their spectra combined with carbon emission; this is the first time such effects have been seen at such early types.

The spectra of 11 of our targets reveal them as SB2 systems, with a possible contribution from a secondary component seen in one other spectrum. From analysis of the RVs estimated from consecutive nights we identified one SB1 system and six candidate RV variables. Eight of these 19 objects are known eclipsing binaries from the OGLE survey (Graczyk et al. 2011). Eight of our other targets were also classified as eclipsing systems by Graczyk et al. (2011), but were undetected as binaries from our spectroscopy.

Using a 3σ threshold compared to the systemic velocity (and excluding the known and candidate binaries), the estimated RVs were used to identify nine RV outliers (Table 8). These are likely to be large-amplitude binaries or runaway stars, and follow-up spectroscopy is required to clarify their nature.


1

In this illustrative calculation we adopt an absolute magnitude for a B0 dwarf of 3.6 (Walborn 1972), a modest extinction of AV= 0.4 mag (from a typical reddening toward the LMC of E(BV)~ 0.13 mag from Massey et al. 1995, and assuming R, the ratio of total-to-selective extinction of 3.1), and a distance modulus to the LMC of 18.5 mag (Pietrzyński et al. 2013).

2

Stock et al. (1976) cross-matched their star 399 to Sk68°100, but Brian Skiff’s updated catalogues (see footnotes to Table 2) match this Sanduleak source to SOI 398 (classified by Stock et al. as A1 Ia).

3

The PoWR models are not tailored to fit our spectra (e.g. in contrast to the observations, the N V lines are predicted in emission for the adopted parameters); our objective was a first investigation of the sensitivity of the C IV emission to different parameters.

4

At such types λ4026 is strongly dominated by He II absorption.

5

For completeness: AAΩ 30 Dor 084, 085, 135, 160, 192, and 419.

Acknowledgments

We are grateful to Nolan Walborn for his careful reading of the draft manuscript, and for his thoughts regarding the classification of the early-type spectra. We also thank the referee for their constructive comments, and Russell Cannon and Gary Da Costa for their help with the observations.

References

Online material

Table 2

Observational parameters of target stars.

Table 3

MCPS photometry of target stars (Zaritsky et al. 2004).

Table 4

Comparison of AAOmega classifications with those available in the literature.

All Tables

Table 1

AAOmega spectrograph settings used.

Table 5

Rest wavelengths used for the radial-velocity estimates.

Table 6

Summary of known binaries and candidate radial-velocity variables in the AAOmega sample.

Table 7

Mean radial velocities and dispersions for the AAΩ sample for both nights ( and , respectively), excluding known and candidate binaries (see Sect. 4.1) and stars with outlying velocities (Table 8).

Table 8

Stars with outlying radial velocities, which are candidate large-amplitude binaries or runaway stars.

Table 2

Observational parameters of target stars.

Table 3

MCPS photometry of target stars (Zaritsky et al. 2004).

Table 4

Comparison of AAOmega classifications with those available in the literature.

All Figures

thumbnail Fig. 1

Location of our AAOmega targets overlaid on a blue-optical image from the Digitized Sky Survey (DSS). The ten early-type stars (with classifications of O4 or earlier) discussed in Sect. 3 are highlighted in red.

Open with DEXTER
In the text
thumbnail Fig. 2

AAOmega spectra of the ten early O-type stars in the sample (with each spectrum smoothed by a 5-pixel median filter for clarity and offset by 0.5 continuum units). Absorption lines identified in the spectrum of AAΩ 30 Dor 364 are: He II λλ4026, 4200, 4542, 4686; He I λ4471; N V λλ4604, 4620. The emission lines identified in the spectra are, in order of increasing wavelength: N IV λ4058; Si IV λλ4089, 4116; N III λλ4634-40-42; C IV λ4658. Broad absorption from the λ4430 diffuse interstellar band can be seen in some spectra (e.g. AAΩ 30 Dor 254).

Open with DEXTER
In the text
thumbnail Fig. 3

λλ4500-4725 range of three early-type AAOmega spectra, illustrating the large variations in N V λλ4604-20 absorption and the presence of C IV λ4658 (and O IV λ4632) emission.

Open with DEXTER
In the text
thumbnail Fig. 4

Synthetic PoWR spectra (black line) adopting parameters for AAΩ 30 Dor 254 from Evans et al. (2010). The spectra plotted in red, moving from the top to bottom, are models which vary the mass-loss rate (Mdot), terminal velocity (v) clumping factor (fcl), microturbulence (ξ), iron abundance (XFe, to investigate possible blanketing effects), and carbon abundance (XC), as detailed in Sect. 3.1. The lines identified in the lower panel are He II λλ4542, 4686; N V λλ4604-20, and C IV λ4658.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.