Free Access
Issue
A&A
Volume 551, March 2013
Article Number A33
Number of page(s) 13
Section Stellar atmospheres
DOI https://doi.org/10.1051/0004-6361/201220721
Published online 14 February 2013

© ESO, 2013

1. Introduction

During the last years a gradually increasing number of O, early B-type, and WR stars have been investigated for magnetic fields, and as a result, about a dozen magnetic O-type and two dozens of early B-type stars are presently known (e.g., Donati et al. 2006; Hubrig et al. 2008; 2011a,b; Martins et al. 2010; Wade et al. 2011; 2012a,b). The recent detections of magnetic fields in massive stars generate a strong motivation to study the correlations between evolutionary state, rotation velocity, and surface composition (Brott et al. 2011; Potter et al. 2012), and to understand the origin and the role of magnetic fields in massive stars (Langer 2012).

The present study is the continuation of our magnetic surveys with FORS 1/2 at the VLT of a sample of massive O and B-type stars in open clusters at different ages and in the field started already in 2005 (e.g., Hubrig et al. 2008; 2011b). In 2011 we presented the results of 41 spectropolarimetric FORS 2 observations of 36 massive stars. Among the observed sample, roughly half of the stars were probable members of clusters at different ages (Hubrig et al. 2011b). The new survey is based on four nights of observations with FORS 2 in 2011 May. Furthermore, we present a number of follow-up observations carried out with the high-resolution spectropolarimeters SOFIN mounted at the Nordic Optical Telescope (NOT) and HARPS mounted at the ESO 3.6 m between 2008 and 2011. The information on the occurrence rate of magnetic fields and field strength distribution is critical for answering the principal question of the possible origin of magnetic fields in massive stars.

In the following, we present 67 new measurements of magnetic fields in 30 massive stars. Our observations and the obtained results are described in Sects. 2 and 3, and their discussion is presented in Sect. 4.

2. Observations and results

2.1. FORS 2 observations

Table 1

List of massive stars observed for this programme.

Table 2

Longitudinal magnetic fields measured with FORS 2 in the studied sample.

Table 2

continued.

thumbnail Fig. 1

Normalised FORS 1/2 Stokes I spectra of O-type stars with detected magnetic fields observed from 2005 to 2011. Well known spectral lines are indicated; all Balmer lines from the Balmer jump to Hβ are visible. The spectra were offset from 1 by multiples of 0.5 for clarity.

Open with DEXTER

The spectropolarimetric observations were carried out in 2011 May 4–8 in visitor mode at the European Southern Observatory (ESO) with FORS 2 mounted on the 8-m Antu telescope of the VLT. This multi-mode instrument is equipped with polarisation analysing optics, comprising super-achromatic half-wave and quarter-wave phase retarder plates, and a Wollaston prism with a beam divergence of 22″ in standard resolution mode1. Polarimetric spectra were obtained with the GRISM 600B and the narrowest slit width of 04 to achieve a spectral resolving power of R ~ 2000. The use of the mosaic detector made of blue optimised E2V chips and a pixel size of 15 μm allowed us to cover a large spectral range, from 3250 to 6215 Å, which includes all hydrogen Balmer lines from Hβ to the Balmer jump. The spectral types and the visual magnitudes of the studied stars are listed in Table 1.

A detailed description of the assessment of the longitudinal magnetic-field measurements using FORS 2 was presented in our previous papers (e.g., Hubrig et al. 2004a,b, and references therein). The mean longitudinal magnetic field, ⟨Bz⟩, was derived using (1)where V is the Stokes parameter that measures the circular polarisation, I is the intensity in the unpolarised spectrum, geff is the effective Landé factor, e is the electron charge, λ is the wavelength, me the electron mass, c the speed of light, dI/dλ is the derivative of Stokes I, and ⟨Bz⟩ is the mean longitudinal magnetic field. Errors are determined in the original spectra from photon statistics and propagated to the resulting V/I spectra. The error of ⟨Bz⟩ is obtained from the formal error of the error-weighted least-squares fit.

Longitudinal magnetic fields were measured in two ways: using only the absorption hydrogen Balmer lines or the entire spectrum including all available absorption lines. The lines that show evidence of emission were not used to determine the magnetic field strength. In addition, a rectification was carried out for a few massive stars with continuum slopes detected in their Stokes V spectra. The feasibility of longitudinal magnetic field measurements in massive stars using low-resolution spectropolarimetric observations has been demonstrated by previous studies of O and B-type stars (e.g., Hubrig et al. 2006; 2008; 2009a; 2011a,b).

Ten targets were observed once, 13 were observed two or three times. HD 148937 was observed four times to assess the magnetic field variability over the rotation cycle. Apart from this star, which has a rotation period of seven days (Nazé et al. 2010), no exact rotation periods are known for the other stars in our sample. The star ζ Oph was observed nine times with the aim to search for magnetic/rotational periodicity (see Sect. 3).

The results of our magnetic field measurements are presented in Table 2. In the first two columns, we provide the star names and the modified Julian dates at the middle of the exposures, followed by the signal-to-noise ratio (S/N) in the individual spectra in Col. 3. In Cols. 4 and 5, we present the longitudinal magnetic field ⟨Bzall using the whole spectrum, and the longitudinal magnetic field ⟨Bzhyd using only the hydrogen lines. All quoted errors are 1σ uncertainties. In Col. 6 we indicate the wavelength regions used for the magnetic field measurements. The spectral regions containing the interstellar lines of the Ca ii doublet are usually not used in the measurements. In Col. 7 we list the rotation phases for the Of?p star HD 148937, which is the only star in our FORS 2 observations with a known period of about seven days. Finally, in the last column we identify new detections by ND, previous detections by PD and confirmed detections by CD.

In our sample of 25 stars observed with FORS 2 in 2011 May the presence of a magnetic field is confirmed in six stars previously observed with FORS 1/2 (Hubrig et al. 2008; 2011b,c): CPD−28 2561, CPD−47 2963, HD 93843, HD 148937, ζ Oph, and HD 328856. New magnetic field detections were achieved in five stars: HD 92206c, HD 93521, HD 93632, CPD−46 8221, and HD 157857. Although previous, mostly single observations of the five O-type stars HD 130298, HD 153426, HD 153919, HD 154643, and HD 156154 indicated the possible presence of weak magnetic fields (Hubrig et al. 2011b), our recent observations do not reveal magnetic fields at a significance level of 3σ in these stars. We note that these recent observations have been carried out on single epochs for four out of five stars, only HD 130298 was observed on two different nights. Since the rotation periods for them are unknown, the non-detection of their magnetic fields can probably be explained by the strong dependence of the longitudinal magnetic field on the rotational aspect.

The strongest magnetic fields are detected in the two Of?p stars CPD−28 2561 and HD 148937. Walborn (1973) introduced the Of?p category for massive O stars displaying recurrent spectral variations in certain spectral lines, sharp emission or P Cygni profiles in He i and the Balmer lines, and strong C iii emission lines around 4650 Å. Only five Galactic Of?p stars are presently known: HD 108, NGC 1624-2, CPD−28 2561, HD 148937, and HD 191612 (Walborn et al. 2010), and all of them show evidence for the presence of magnetic fields (Martin et al. 2010; Wade et al. 2012a; Hubrig et al. 2008; 2011b; Donati et al. 2006). The record holder is the faintest star in the sample of Of?p stars, NGC 1624-2, for which a longitudinal magnetic field of the order of 5 kG was recently detected by Wade et al. (2012a).

Besides the small group of Of?p stars, our measurements indicate that the presence of magnetic fields can be expected in stars of different classification categories and at different evolutionary stages, from young main-sequence stars up to the supergiant stages. The different spectral appearance of stars with confirmed and new detections is illustrated in Fig. 1 where we present normalised FORS 1/2 spectra in integral light. Among the whole sample, the two Of?p stars CPD−28 2561 and HD 148937 exhibit the most conspicuous spectral variability clearly visible in Fig. 1, showing the spectra of these stars obtained at two different epochs. To calculate the rotation phase of the observed spectra of HD 148937, we used the rotation period of 7.03 d determined by Nazé et al. (2008). The stars with confirmed and new magnetic field detections are discussed in more detail in Sect. 3.

2.2. SOFIN and HARPS observations

Four O-type stars, HD 36879, 15 Mon, 9 Sgr, and HD 191612, previously observed with FORS 1 and SOFIN (Hubrig et al. 2008; 2009a; 2010), were followed-up by nine additional SOFIN spectropolarimetric observations from 2008 September and 2010 July. Spectral classifications of these stars are presented in Table 1. Among these stars, only for the Of?p star HD 191612 was the presence of a magnetic field of the order of a few hundred Gauss previously reported by Donati et al. (2006). We used the low-resolution camera (R ≈ 30   000) of the echelle spectrograph SOFIN (Tuominen et al. 1999) mounted at the Cassegrain focus of the NOT. With the 2K Loral CCD detector we registered 40 echelle orders partially covering the range from 3500 to 10 000 Å with a length of the spectral orders of about 140 Å at 5500 Å. The polarimeter is located in front of the entrance slit of the spectrograph and consists of a fixed calcite beam splitter aligned along the slit and a rotating super-achromatic quarter-wave plate. Two spectra polarised in opposite sense are recorded simultaneously for each echelle order providing sufficient separation by the cross-dispersion prism. Two to four sub-exposures with the quarter-wave plate angles separated by 90° were used to derive circularly polarised spectra. A detailed description of the SOFIN spectropolarimeter and its polarimetric data reduction is given in Ilyin (2012).

The spectra were reduced with the 4A software package (Ilyin 2000). Bias subtraction, master flat-field correction, scattered light subtraction, and weighted extraction of spectral orders comprise the standard steps of spectrum processing. A ThAr spectral lamp is used for wavelength calibration, taken before and after each target exposure to minimise temporal variations in the spectrograph. The typical S/N for observations of 9 Sgr is between 280 and 380, while for HD 36879 and HD 191612 it is about 200. The highest S/N of 440 was achieved for 15 Mon.

Table 3

Magnetic field measurements of five O-type stars using SOFIN and HARPS observations.

In addition, the two O-type stars, HD 151804 and 9 Sgr, previously analysed using FORS 1 data (Hubrig et al. 2008) were observed on 2011 May 25 and 26 with a S/N of about 280 with the HARPSpol polarimeter (Snik et al. 2011) feeding the HARPS spectrometer (Mayor et al. 2003) at the ESO 3.6 m telescope on La Silla within the framework of the programme 187.D-0917(A). This HARPS spectropolarimetric material became recently publically available in the ESO archive. The HARPS archive spectra cover the wavelength range 3780–6913 Å, with a small gap around 5300 Å. Each HARPSpol observation is usually split into four to eight sub-exposures, obtained with four different orientations of the quarter-wave retarder plate relative to the beam splitter of the circular polarimeter. The reduction was performed using the HARPS data reduction software available at the ESO headquarters in Germany. The Stokes I and V parameters were derived following the ratio method described by Donati et al. (1997), ensuring in particular that all spurious signatures are removed at first order. Null polarisation spectra (labelled with n in Table 3) were calculated by combining the sub-exposures in such a way that the polarisation cancels out, allowing us to verify that no spurious signals are present in the data.

HARPS and SOFIN spectra have been measured using the moment technique previously introduced by Mathys who discussed at length the diagnostic potential of high-resolution circularly polarised spectra using this technique in numerous papers (e.g., Mathys 1993; 1995a,b). Wavelength shifts between right- and left-hand side circularly polarised spectra are interpreted in terms of a longitudinal magnetic field ⟨Bz⟩. ⟨Bz⟩ is obtained by performing a least-squares fit, forced through the origin, of the wavelength shifts λR − λL, as a function of . In this process, the individual measurements are weighted by the respective 1/σ2(λR − λL). The uncertainty of the derived value of the longitudinal magnetic field is estimated by its standard error σ(⟨Bz⟩), derived from the least-squares analysis (e.g., Mathys & Hubrig 1995). The results of the longitudinal magnetic field measurements using mostly He i and He ii lines are presented in Table 3.

The high-resolution spectropolarimetric observations confirm our previous detections of magnetic fields in the stars HD 36879 and 9 Sgr. The magnetic nature of the slowly rotating Of?p star HD 191612 (P = 538 d) was for the first time reported by Donati et al. (2006) who observed it over four nights in 2005 June using the ESPaDOnS spectropolarimeter installed on the 3.6-m Canada-France-Hawaii Telescope and announced the detection of a mean longitudinal field  ⟨ Bz ⟩  = −220 ± 38 G. Hubrig et al. (2010) reported on a single measurement of the magnetic field of HD 191612,  ⟨ Bz ⟩  = 450 ± 153 G, using SOFIN observations acquired in 2008 September at rotational phase 0.43. Wade et al. (2011) presented 13 new field measurements of this star demonstrating that the magnetic data can be modelled as a periodic, sinusoidal signal with a period of 538 d inferred from spectroscopy. The authors mention in their work that the SOFIN observation of the longitudinal field of HD 191612,  ⟨ Bz ⟩  = 450 ± 153 G, does not agree very well with the observed field variation, which corresponds to (an essentially) consistently negative longitudinal magnetic field. To check the consistency of our measurements we re-analysed the previous SOFIN observations using an extended list of blend free spectral lines with known atomic parameters. The obtained results,  ⟨ Bz ⟩  = 357    ±    58 G, indicates, however, that contrary to the results of Wade et al. (2011), the magnetic field of HD 191612 is indeed rather strongly positive in the phase 0.43. Clearly, more magnetic field observations of this star are needed to properly characterise the variation of the longitudinal magnetic field and the magnetic field geometry. In Fig. 2 we present our SOFIN observations together with those published by Wade et al. (2011).

thumbnail Fig. 2

Longitudinal magnetic field variation of the Of?p star HD 191612 according to the 537.6 d period determined by Howarth et al. (2007). Red symbols correspond to ESPaDOnS observations, while blue symbols are our SOFIN measurements (Hubrig et al. 2010, and Table 3). Note that the measurement errors for both ESPaDOnS and SOFIN observations are of similar order.

Open with DEXTER

The supergiant HD 151804 was previously observed three times in 2005 with FORS 1 (Hubrig et al. 2008). No field detection at a significance level of 3σ was achieved indicating that the field should be rather weak if present at all. Our measurement  ⟨ Bz ⟩  = −104 ± 38 G based on HARPSpol observations confirms these previous results.

3. Results for individual targets with detected magnetic fields

3.1. HD 36879

thumbnail Fig. 3

Variable line profiles in the SOFIN spectra of HD 36879 obtained at two different epochs. Left panel: spectral variability of the Fe iii line profile at λ6048 and the blend Fe iii + He ii at λ6078. Right panel: spectral variability of line profiles presenting blends of the C iii and O ii lines at λ4647 and λ4660.

Open with DEXTER

thumbnail Fig. 4

Line profiles of He i in the spectra of HD 36879 obtained at two different epochs. From top left to bottom right we show the lines He i λ4713, λ4922, λ5015, and λ5876.

Open with DEXTER

A magnetic field at a 3.5σ significance level in HD 36879 was for the first time detected in 2007 using FORS 1 observations (Hubrig et al. 2008). Our most recent high-resolution SOFIN observations in 2009 December confirm the previous detection and reveal a change of the field polarity. Walborn (2006) reports that this star shows unexplained spectral peculiarities and/or variations, among them peculiar narrow, variable Si iv emission lines at λ1394 and λ1403 (e.g. Walborn & Panek 1984). Distinct spectral variability is also discovered in Fe iii and the blended lines C iii/O ii using our two high-resolution SOFIN spectra obtained on two different epochs. The variations of Fe iii and C iii/O ii line profiles are presented in Fig. 3. Interestingly, He i lines appear slightly split. This appearance presented in Fig. 4 could possibly indicate the presence of He spots on the stellar surface.

3.2. 15 Mon

Hubrig et al. (2011b) showed that this star is a probable member of the young, rich open cluster NGC 2264 located in the Monoceros OB 1 association. From the spatial distribution of Hα emission stars, Sung et al. (2008) have concluded that the young open cluster NGC 2264 can be subdivided into three star forming regions (SFRs): two active SFRs – the region around the bright emission nebula in the southwest of the brightest star S Mon and the region to the north of the famous cone nebula – and surrounding these an elliptical-shaped halo region. The study of Teixeira et al. (2012) indicates that NGC 2264 has undergone more than one star-forming event, where the unembedded, insubstantial inner disk population and embedded thick inner disk population appear to have formed in separate episodes. Markova et al. (2004) consider 15 Mon as a Galactic O-type star with a mass of 32 M and Teff = 37   500 K. Speckle measurements of 15 Mon indicated the presence of a companion of spectral type O9.5 V n (e.g. Gies et al. 1993). In very high S/N spectra becomes the companion noticeable by very broad underlying components to the strong He i lines. According to Kaper et al. (1996) and Walborn (2006) shows this star distinct peculiarities in the spectra, which could be typical for stars possessing magnetic fields. The X-ray luminosity based on Chandra observations is log    (LX) = 31.82 and the ratio of X-ray to bolometric luminosity log    (LX/Lbol) = −6.63 is slightly above the average (Ramirez et al. 2004; Nazé 2009), as may be expected from a magnetic star. The high-resolution X-ray spectrum of 15 Mon is typical for a star with its spectral type and, apparently, not especially hard (Walborn et al. 2009).

The previously measured magnetic field using the full FORS 1 spectrum coverage was of positive polarity at a 2.6σ significance level (Hubrig et al. 2009a). A weak magnetic field was detected in this star at a significance level of 4.4σ using SOFIN observations. No contribution of the secondary component was detected in our FORS 1 and SOFIN spectra (see also Fig. 1).

3.3. CPD−28 2561

thumbnail Fig. 5

Distinct variability observed in the spectra of CPD−28 2561 in the He i λ4026 (top left), He ii λ4686 (top right), Hβ (bottom left), and He i λ5876 (bottom right) lines.

Open with DEXTER

Our previous observations of the Of?p star CPD−28 2561 enabled us to detect a magnetic field at the 3.1σ level using the whole spectrum and at the 3.2σ level using Balmer lines (Hubrig et al. 2011b). Our new spectropolarimetric observations of this star, recently reported by us in a short IBVS article (Hubrig et al. 2012a) indicate that the magnetic field is variable. Due to the relative faintness of CPD−28 2561 with V = 10.1, it was only scarcely studied in the past. Levato et al. (1988) acquired radial velocities of 35 OB stars with carbon, nitrogen, and oxygen anomalies and found variability of a few emission lines with a probable period of 17 days. Walborn et al. (2010) mentioned that CPD−28 2561 undergoes extreme spectral transformations very similar to those of HD 191612, on a time scale of weeks, inferred from the variable emission intensity of the C iii λλ 4647-4650-4652 triplet. Our spectropolarimetric observations on three consecutive nights in 2011 reveal strong variations in several hydrogen and helium line profiles. The variation of He i λ4026, He i λ5876, and of Hβ on such a short time scale were not reported in the literature before. A few examples of the detected spectral line profile variations are presented in Fig. 5. The contribution of a variable emission component is well visible in three lines. The profile of the He i λ4026 line appears only weakly variable. These variations have an amplitude of about 1–2% of the continuum level and can likely be explained by rotation modulation. The line He ii λ4686 exhibited strong emission during our observations in 2010, which transforms to a much weaker P Cygni line profile with a weak blue shifted absorption in 2011. A contribution of the emission component is clearly seen in the profile of the He i λ5876 line in 2010, but its presence can only be suspected in the variable line depth observed in 2011. A strong emission component is well visible in the line centre of Hβ in 2010. The shape of the line profile of Hβ is similar to the behaviour of the Hα line profile in the spectra of 19 Cep presented in Fig. 13 of Kaper et al. (1997) and in Fig. 1 of Kholtygin et al. (2003).

CPD−28 2561 is an intrinsically bright X-ray source as detected by the Rosat X-ray observatory. Its ratio of X-ray to bolometric luminosity (log LX/Lbol ≈ −6) makes it unusually X-ray bright compared to other O-type stars (Chlebowski 1989; Oskinova 2005). The hardness ratio (HR1 ≈ 0.04 ± 0.4) indicates that its X-ray emission can be harder than is typical in single O-type stars. This provides indirect support for the presence of a magnetic field on the star, strong enough to confine the stellar wind (Babel & Montmerle 1997).

3.4. CPD−47 2963

For the star CPD−47 2963, we confirm our previous 3.1σ detection by measuring all absorption lines in the FORS 2 spectra. The magnetic field of this star is variable and shows a change of polarity over our four-night observing run. According to Walborn et al. (2010), CPD−47 2963 belongs to the Ofc category, which consists of normal spectra with C iii λλ 4647-4650-4652 emission lines of comparable intensity to those of the Of defining lines N iii λλ 4634-4640-4642. The origin of the magnetic field in this star probably differs from those of other magnetic O-type stars, because non-thermal radio emission, which is frequently observed in binary systems with colliding winds, was detected by Benaglia et al. (2001). On the other hand, the membership of CPD−47 2963 to a binary or multiple system has not yet been investigated.

3.5. HD 92206c

The presence of a variable magnetic field up to a 4.4σ significance level in HD 92206c is documented here for the first time. This star is a probable member of the open cluster NGC 3324 located inside of a partial ring of nebulosity northwest of the Carina Nebula. The assessment of the membership using astrometric catalogues with the highest quality kinematic and photometric data currently available is described in Sect. 4. According to the Galactic O Star Catalogue (Maíz-Apellániz et al. 2004) belongs HD 92206c to the triple system HD 92206abc. An analysis of the distribution of the ionised and neutral material associated with the H ii region Gum 31 by Cappa et al. (2008) suggested that the excitation sources of this region are the brightest stars in the open cluster NGC 3324 with the triple system HD 92206abc considered as the brightest cluster member. The detection of protostellar candidates almost coincident in position with the open cluster NGC 3324 indicates that stellar formation is ongoing in this region.

Walborn (1982) reports that the spectrum of HD 92206c displays very strong, broad hydrogen lines, possibly similar to those in the Orion Trapezium cluster, in particular θ1 Ori C, and the appearance of these lines is indicative of extreme youth. A search for a magnetic field was carried out in another member of this cluster, the bright A0-type supergiant HD 92207. According to Hubrig et al. (2012b), this star also possesses a magnetic field.

3.6. HD 93521

A longitudinal magnetic field of the order of 200–300 G is detected in this star at a significance level of 3.2σ. This high Galactic latitude O9.5 Vp star with v sin i = 390 km s-1 (Rauw et al. 2008) is one of the fastest rotators in our sample, along with ζ Oph. It is located at an unusually high Galactic latitude of 62.15 and displays prominent line profile variability in the optical and UV domain (Rauw et al. 2008). In the most recent work by Rauw et al. (2012a), the authors report that He and N are found to be overabundant and suggest that despite some ambiguities on the runaway status of the star it is likely that the star is a Population I massive O-type star that was ejected from the Galactic plane either through dynamical interactions or as the result of a supernova event in a binary system.

HD 93521 is one of the few O-type stars known to exhibit stellar pulsations. In this star, two frequencies of about 2–3 h are detected and interpreted as due to high-degree non-radial pulsation modes (Howarth & Reid 1993; Rauw et al. 2008).

The X-ray observations of this star have been summarised by Rauw et al. (2012a). They concluded that the stellar X-ray spectrum is consistent with a normal late O-type star although with subsolar metallicity. No trace of a compact companion was found in the X-ray data. The ratio of X-ray to bolometric luminosity is log    (LX/Lbol) = −7.1...−7.0 and the stellar X-ray luminosity LX = 5 × 1029 erg s-1 is rather low. Such low X-ray luminosities are also observed in other magnetic stars. Oskinova et al. (2011) showed, that relatively low X-ray luminosities can be explained in the framework of the magnetically confined wind shock model if the stellar wind strength is correctly derived from an optical and UV spectroscopic analysis. Significantly, Rauw et al. (2012a) show that besides a softer plasma component with kT1 = 0.3 keV, which is typical for normal O-stars, a harder component with kT2 = 3 keV is present in the X-ray spectrum of HD 93521. Recently, Ignace et al. (2013) showed that such a hard component may serve as an indicator for the presence of a stellar magnetic field.

3.7. HD 93632 and HD 93843

A weak negative magnetic field in HD 93632, just at a significance level of 3σ was detected in one out of two FORS 2 observations. Walborn (1973) reported that both HD 93632 and HD 93843 appear slightly variable in their spectral types. Interestingly, also HD 93843, similar to HD 93632, exhibits a weak magnetic field. It’s presence is confirmed in our new observations. According to Walborn et al. (2010), the star HD 93843 belongs to the Ofc category. Prinja et al. (1998) monitored the stellar wind of this star using IUE time series. They identified systematic changes in the absorption troughs of the Si iv and N v resonance lines with a repeatability of wind structures with a period of 7.1 days. The authors suggested at that time that a magnetic field might help to explain the cyclical wind perturbation. Both stars have been observed during the Chandra X-ray survey of the Carina Nebula (Townsley et al. 2011). For HD 93632 is the X-ray luminosity log LX = 31.7. The ratio of X-ray to bolometric luminosity is small, only log    (LX/Lbol) = −7.7. A one-temperature fit to the X-ray spectrum shows a slightly higher than average temperature kT = 0.56 (Nazé et al. 2011). For HD 93843, the X-ray luminosity is log LX = 32.1. The ratio of X-ray to bolometric luminosity is standard, log    (LX/Lbol) = −7.1. A one-temperature fit to the X-ray spectrum derives a spectral temperature kT = 0.3, which is lower than in HD 93632 (Nazé et al. 2011).

3.8. HD 148937

thumbnail Fig. 6

Longitudinal magnetic field variation of the Of?p star HD 148937 according to the 7.032 d period determined by Nazé et al. (2010). Red symbols correspond to the ESPaDOnS observations (Wade et al. 2012b), while green symbols are our FORS 1 and FORS 2 measurements (Hubrig et al. 2008; 2011b and present paper). Note that the measurement errors for both ESPaDOnS and FORS 1/2 observations are of similar order. One more measurement (not shown in this figure) was obtained in 2010 on May 22, but ordinary and extraordinary beams in the FORS 2 setup were overlapping in this exposure due to problems with slitlets.

Open with DEXTER

The first detection of a mean longitudinal magnetic field  ⟨ Bz ⟩  = −254 ± 81 G in the Of?p star HD 148937 using FORS 1 at the VLT was reported by Hubrig et al. (2008). An extensive multiwavelength study of HD 148937 was carried out by Nazé et al. (2008), who detected small-scale variations of He ii λ4686 and the Balmer lines with a period of seven days and an overabundance of nitrogen by a factor of four compared to the Sun. The periodicity of spectral variations in hydrogen and helium lines was re-confirmed using additional higher resolution spectroscopic material indicating the similarity to the other Of?p stars HD 108 and HD 191612 (Nazé et al. 2010). Our new and old observations of HD 148937 are presented in Fig. 6 together with the ESPaDOnS observations obtained at the CFHT by Wade et al. (2012b). This figure demonstrates the excellent potential of FORS 2 for the detection and investigation of magnetic fields in massive stars: While an exposure time of 21.5 h at the CFHT was necessary to obtain seven binned measurements, the exposure time for FORS 2 observations accounted only for two to four minutes and only 2.3 h were used for our observations at six different epochs including telescope presets and the usual overheads for readout time and retarder waveplate rotation.

3.9. ζ Oph

thumbnail Fig. 7

Phase diagram for the best sinusoidal fit corresponding to the period of 0.8 days for the longitudinal magnetic field measurements using hydrogen lines for ζ Oph.

Open with DEXTER

thumbnail Fig. 8

Phase diagram for the best sinusoidal fit corresponding to the period of 1.3 days for the longitudinal magnetic field measurements using the whole spectrum for ζ Oph.

Open with DEXTER

The star ζ Ophiuchi is a well-known rapidly rotating runaway star with extremely interesting characteristics. This star exhibits a wonderful bow shock as evident from IR Spitzer IRAC observations and undergoes episodic mass loss seen as emission in Hα. It is possible that ζ Oph rotates with almost break-up velocity with v sin i = 400 km s-1 (Kambe et al. 1993). ζ Oph is the second target of our sample with detected non-radial pulsations with periods of hours, first discovered from ground-based observations (e.g. Kambe et al. 1997). Later, the star was monitored by the MOST satellite, which revealed at least a dozen significant pulsation frequencies between 1 and 10 d-1, typical of β Cep-like pulsation modes (Walker et al. 2005).

The first detection of a mean longitudinal magnetic field in this star was announced by Hubrig et al. (2011c). Nine additional FORS 2 spectropolarimetric observations showing a change of polarity have been obtained over all four observing nights. The rotation period of this star is still uncertain: Oskinova et al. (2001) studied ASCA observations that covered just the expected rotation period of the star of about 0.8–1.0 d. A X-ray flux variability with  ~20% amplitude in the ASCA passband (0.5–10 keV) with a period of 0.77 d was reported. The X-ray spectra of ζ Oph are not especially hard as found from Chandra HETGS and Suzaku data (Waldron & Cassinelli 2007; Walborn et al. 2009). The variation period is also not well-defined in our observations. The resulting periodograms for the magnetic field measurements using hydrogen lines shows a dominating peak corresponding to a period of about 0.8 d, while the measurements using the whole spectrum indicate a rotation period of about 1.3 d. In Fig. 7 we present the magnetic field variation using hydrogen lines phased with the period of 0.8 days, while Fig. 8 refers to magnetic field variations determined from all lines phased with 1.3 days. Interestingly, just a few months ago Pollmann (2012) presented equivalent width changes of He i λ6678 with a period of 0.643 d, which is roughly half of the magnetic period of 1.3 d determined using the whole spectral region. For the rotation period of  ~1.3 d the equivalent width of the He i line would display a double wave. Such a behaviour of He i lines is frequently found in He variable early-type Bp stars (e.g. Briquet et al. 2004). Clearly, more measurements are needed to determine the true periodicity of the magnetic field in this star.

thumbnail Fig. 9

Combined IR Spitzer IRAC (3.6 μm blue, 4.5 μm green, 8.0 μm red) image of the bow shock around Vela X-1. Archival data have been used (AOR 17858304, PI: R. Iping). Image size is 5′ × 5′.

Open with DEXTER

A bow-shock nebula is also present close to the high mass X-ray binary Vela X-1, consisting of a pulsar in an eccentric (e = 0.09,Porb = 8.96 d) orbit around the here studied B0.5 supergiant star HD 77581. In Fig. 9 we present an image of the bow-shock nebula based on archival Spitzer IRAC maps (AOR 17858304, PI R. Iping). The mass-loss rate ( ≈ 10-6   M yr-1) for Vela X-1 is well constrained from its accretion powered X-ray luminosity (e.g., Fürst et al. 2010; Oskinova et al. 2012). Contrary to our study of ζ Oph, no magnetic field detection at a significance level of 3σ has been achieved in Vela X-1 during our observing run. However, we should keep in mind that the spectral behaviour of HD 77581 is much more complex than that of ζ Oph due to the presence of bumps and wiggles in the line profiles, and an impact of tidal effects producing orbital phase-dependent variations in the line profiles leading to asymmetries such as extended blue or red wings (Koenigsberger et al. 2012). Furthermore, Kreykenbohm et al. (2008) detected flaring activity and temporary quasi-periodic oscillations in INTEGRAL X-ray observations.

3.10. HD 328856

Our new observations confirm the presence of a magnetic field in this star. As we already showed in our previous work on the basis of photometric membership probability, HD 328856 is a member of the compact open cluster Hogg 22 in the Ara region at an age of 5 Myr and a distance of about 1300 pc (Hubrig et al. 2011b). On the other hand, its proper motions indicate that HD 328856 is not fully co-moving with the other cluster members, deviating from the cluster mean proper motion by  ~2σ (for more details of the membership probabilities, see Kharchenko et al. 2004).

3.11. CPD−46 8221

A search for the presence of a magnetic field in this star was carried out for the first time. A weak mean longitudinal field was detected during the first observing night at a significance level of 3.1σ. As we show below in Sect. 4, this star has a high kinematical membership probability in the open cluster Hogg 22.

3.12. HD 157857

Spectropolarimetric observations of this star have been conducted for the first time. Our measurements on the second observing night carried out on hydrogen lines resulted in a field detection at a significance level of 3.7σ. Schilbach & Roeser (2008) retraced the orbit of HD 157857 in the Galactic potential and concluded that it was ejected from the open cluster NGC 6611 about 3.8 Myr ago with a peculiar velocity of 113.5 km s-1.

3.13. HD 164794

All four SOFIN and one HARPS high-resolution spectropolarimetric observations confirm our previous detection of the longitudinal magnetic field in this star. The field appears to be variable with negative and positive extreme values of −265 G and  + 242 G. HD 164794 belongs to the young cluster NGC 6530 and according to the Galactic O Star Catalogue (Maíz-Apellániz et al. 2004), has a classification O4 V ((f)), exhibiting weak N iii emission and strong He ii λ4686 absorption. It is a well-known non-thermal radio emitter and, according to van Loo et al. (2006), the most likely mechanism is synchrotron emission from colliding winds, implying that all O stars with non-thermal radio emission should be members of binary or multiple systems. Hints of a wind-wind interaction were indeed detected in the X-ray domain (Rauw et al. 2002). A long-term study of its binary nature and spectrum variability has been recently presented by Rauw et al. (2012b) who derived an orbital solution and an orbital period of 8.6 yr.

3.14. HD 191612

Our SOFIN spectropolarimetric observations of this star have already been discussed in Sect. 2.2. HD 191612 is currently one of the best studied targets in the optical and X-ray domains among the five currently known Galactic Of?p stars. Due to the above mentioned discrepancies in the magnetic field measurements around the positive field extremum, additional magnetic field observations are necessary for proper magnetic field geometry modelling.

4. Discussion

Whereas magnetic fields are ubiquitous in low mass stars due to their convective envelopes, and found in 10...20% of all intermediate mass stars, they were unexpected in O stars until just a few years ago. The first O star for which a magnetic field was detected is the spectroscopically variable star θ1 Ori C (Donati et al. 2002). However, with ever more magnetic O stars being identified in recent years (e.g., Hubrig et al. 2011b; Martins et al. 2012), it appears likely that the as yet small number of magnetic O stars is just a consequence of an observational bias, due to the low number of spectral lines in these objects. Indeed, there may be more reasons to find magnetic fields in O stars than in A and B stars. Fossil fields might appear with similar frequencies on both groups. If close binary effects as mass transfer or stellar merger would produce large scale stable surface fields in O stars (Ferrario et al. 2009; Langer 2012), then about one quarter of them would be expected to be magnetic due to their very high close binary fraction (Sana et al. 2012). Furthermore, their proximity to the Eddington limit produces convection in the envelopes of O type stars, which could also produce surface magnetic fields (Cantiello et al. 2009).

With the present work we gradually increase the sample of massive stars with determined magnetic field strengths. It is the aim of our survey to provide new clues about the origin and detectability of the magnetic fields of massive stars. Bagnulo et al. (2012) used the ESO FORS 1 pipeline to reduce the full content of the FORS 1 archive. They claim that very small instrument flexures, negligible in most of the instrument applications, may be responsible for some spurious magnetic field detections. We note, however, that out of the five Of?p stars, two were detected to be magnetic for the first time by us (Hubrig et al. 2008; 2011b). The excellent agreement of our FORS 1 and FORS 2 data with ESPaDOnS observations by Wade et al. (2012b) of HD 148937 can be seen in Fig. 6. Bagnulo et al. (2012) could not detect the magnetic field in 9 Sgr (=HD 164794), while our high-resolution observations of this star with both SOFIN and HARPS fully confirm our earlier result (Hubrig et al. 2008). Also, high-resolution observations with ESPaDOnS (Petit et al. 2013) showed a weak magnetic field in 40 Tau (=HD 25558), which was already detected by Hubrig et al. (2009b), but not found by Bagnulo et al. (2012) from the same data. Overall, it is not clear why Bagnulo et al. (2012) are not able to find magnetic-fields found by other groups with FORS 1/2 and other instruments. These are only a few examples that show the need for confirmations of FORS 1 magnetic-field measurements with high spectral resolution polarimeters. In the following, we discuss the role of magnetic stars in star clusters, magnetic runaway stars, and indirect magnetic field indicators in massive stars.

4.1. Magnetic stars in young clusters

Table 4

Probable members in open clusters.

Of the 30 massive stars analysed in this paper, eleven stars have been found to be related to open clusters of different ages. The data on the cluster membership of these probable cluster O-type stars are presented in Table 4. As database for the compilation of Table 4, we used the All-sky Compiled Catalogue of 2.5 million stars (ASCC-2.5, 3rd version) of Kharchenko & Roeser (2009). The kinematic and photometric probabilities for cluster membership presented in Cols. 3 and 4 were calculated according to the procedures described by Kharchenko et al. (2004). For calculating the kinematic membership probability, only proper motion information was used. Since the star HD 92206c in the open cluster NGC 3324 belongs to the triple system HD 92206abc, we present the proper motions of all triple system members in the same table. Largely deviating proper motions of the cluster member HD 92206b in comparison to the average cluster proper motions are very likely caused by the orbital motion in the pair HD 92206ab. Interestingly, not only the cluster member HD 92206c shows the presence of a magnetic field. According to Hubrig et al. (2012b), also another cluster member, the supergiant HD 92207, possesses a magnetic field of the order of a few hundred Gauss. Clearly, the cluster NGC 3324 should be included in future magnetic field surveys of massive stars to better understand the role of the environment on the generation of their magnetic fields.

As we already mentioned in our previous work, according to Schilbach & Roeser (2008), the star HD 153919 was probably ejected from the cluster NGC 6231 at an age of about 6.5 Myr. At present, HD 153919 is located 4.2° from the centre of NGC 6231. Considering the proper motions of HD 153919 given by Kharchenko & Roeser (2009; μα =  +1.42 ± 1.17 mas/yr, μδ =  +4.84 ± 0.87 mas/yr) and those of the cluster by Kharchenko et al. (2005a; −0.39 ± 0.52 mas/yr, −1.99 ± 0.38 mas/yr), then 2.15 Myr ago the star was at 0.55 ± 0.95° from the cluster centre.

Among the stars presented in Table 4, only five stars show the presence of magnetic fields, HD 47839, HD 92206c, HD 328856, CPD−46 8221, and HD 164794. Two of them, HD 328856 and CPD−46 8221, belong to the open cluster Hogg 22, which appears as a second promising cluster for investigating the magnetic nature of cluster members. Although we deal here with small number statistics, we cannot refrain from mentioning that these five stars with detected magnetic fields belong to the youngest clusters in our small sample.

4.2. Magnetic massive runaway stars

Magnetic fields are detected in stars of spectral types, with different kinematical status, and at different evolutionary stages. Clearly, the detection of magnetic fields in all five Galactic Of?p stars implies a tight relation between the spectral characteristics of the Of?p star group and the presence of a magnetic field. Remarkably, apart from the Of?p star NGC 1624-2, no other Of?p star belongs to an open cluster or association. The analysis of the stellar content of the cluster NGC 1624 by Jose et al. (2011) indicated that NGC 1624-2 is the main ionising source of the region, and the maximum post-main-sequence age of the cluster is estimated as  ~4 Myr.

Furthermore, recent studies of the evolutionary state of several magnetic massive stars indicate that some of them evolved significantly or have the status of a runaway star (e.g., Martins et al. 2010; Hubrig et al. 2011b,d). In the literature, the runaway status or candidate runaway status (i.e., if stars possess high space velocities determined through proper motion and/or radial velocity measurements) is assigned to several magnetic massive stars, such as ζ Oph (Blaauw 1952), HD 93521 (Gies 1987), HD 157857 (Schilbach & Roeser 2008), HD 57682 (Comeron et al. 1998), and the three Of?p stars HD 108, HD 148937, HD 191612 (Hubrig et al. 2011d). It is known that some of the massive stars located far from star clusters and star-forming regions are runaways. They were likely formed in embedded clusters and then ejected into the field because of dynamical few-body interactions or binary-supernova explosions (Eldridge et al. 2011). Although the group of field O stars whose runaway status is difficult to prove does exist, the recent study of Gvaramadze et al. (2012) showed no significant evidence in support of the in-situ proposal for the origin of isolated massive stars.

In some cases, even if the star is considered as an open cluster member, a careful study of their kinematic characteristics indicates that it is not fully co-moving with the other cluster members. As an example, the space motion of the magnetic O-type star θ1 Ori C was studied by van Altena et al. (1988), who reported that θ1 Ori C is moving at 4.8  ±  0.5 km s-1 towards position angle 142° and that this velocity is significantly larger than the dispersion value of 1.5  ±  0.5 km s-1 found for the other cluster members. The results of the radial velocity study of Stahl et al. (2008) indicate that this star is moving rapidly away from the Orion Molecular Cloud and its host cluster. Another example is the kinematical status of the magnetic O-type star HD 328856, which is on the basis of the photometric membership probability a member of the compact open cluster Hogg 22 in the Ara region at an age of 5 Myr and a distance of about 1300 pc. However, its proper motions indicate that HD 328856 is not fully co-moving with the other cluster members, deviating from the cluster mean proper motion by  ~2σ (Hubrig et al. 2011b).

Among our studied runaway stars, HD 93521 and ζ Oph are magnetic fast rotators with β Cep-type pulsations. These stars are thus particularly interesting to calibrate stellar rotation models including magnetic fields in the hot upper part of the Hertzsprung-Russell diagram. Variability linked to stellar pulsation has never been searched for or found in any of our other sample targets. In fact, there are only a few O-type stars for which convincing evidence of non-radial pulsations with periods of a few hours has been detected from ground-based observations. Besides HD 93521 and ζ Oph, there are ξ Per and λ Cep (de Jong et al. 1999). For such stars, the photometric amplitudes are below mmag level, which are very difficult to detect from the ground. It may explain why only few pulsating O-type stars have been identified so far and mostly spectroscopically. Recently, the very high-precision photometry of the CoRoT satellite led to the detection of stellar pulsation in three other O-type objects that belong to the young open cluster NGC 2244 inside the Rosette nebula and to the surrounding association Mon OB2. β Cep-like pulsation modes have been found in HD 46202 (Briquet et al. 2011) and in HD 47129 (Mahy et al. 2011) while unexpected modes of stochastic nature have been detected in HD 46149 (Degroote et al. 2010). Efforts to discover magnetic massive pulsators are worthwhile since their study provides a unique opportunity to probe the impact of a magnetic field on the physics of mixing inside stellar interiors.

4.3. Indirect B-field indicators

One of the major characteristics of magnetic Ap and Bp stars is the spectral variability due to the presence of chemical spots showing a certain symmetry with respect to the magnetic field geometry. Although spectral variability is present in all discovered magnetic massive stars, the interpretation of line profile variations seems not to be as simple as for Ap and Bp stars: the behaviour of line profiles of different ions of a given element does not clearly correspond to the abundance variations. Synthetic spectra calculations of the line profile variations in massive stars are furthermore more troublesome since different spectral lines are to a different degree sensitive to the stellar wind and to non-LTE effects. In any case, a detection of line profile variations can certainly be used as an indirect indicator for the possible presence of a magnetic field. Another aspect that could hint at the presence of a magnetic field in massive stars is the fact that a number of magnetic OB-type stars show nitrogen enrichment (e.g. Morel et al. 2008).

It is not clear yet how to use the X-ray emission as an indirect indicator for the presence of magnetic fields. Anomalous X-ray emission can be expected from a massive star with a magnetic field strong enough to regulate the stellar wind motion. The relative importance of magnetic fields in gases is described by the plasma-βp parameter with βp = 8πp/B2, where p is the gas pressure. The gas is magnetically dominated when βp < 1. For supersonic flows, such as stellar winds, the ram pressure exceeds the gas pressure and the dynamical importance of a magnetic field is defined by the ratio of wind kinetic to magnetic energy density: when this ratio is less than one, the magnetic field dominates the wind bulk motion.

Oskinova et al. (2011) investigated the X-ray emission and wind properties among the complete sample of known magnetic early B-type stars with available X-ray observations. The authors have found that the X-ray emission from magnetic B-type stars is diverse. While some stars display hard variable X-rays, others are rather soft sources. The X-ray luminosities differ among otherwise similar Bp stars by more than two orders of magnitude. Among magnetic O-type stars, the situation is similar. While some of them are bright, hard, variable X-ray emission as predicted by the magnetically confined wind-shock scenario (e.g. Schulz et al. 2000), others are relatively soft and faint X-ray sources (Nazé et al. 2010).

Thus, the analysis of X-ray observations of magnetic OB stars shows that some of them are lacking strong and hard X-ray emission, or even are X-ray dark. Therefore, while a strong hard X-ray emission serves as a good indicator for the potential presence of a strong stellar magnetic field, X-ray strong stars represent only a sub-class of all magnetic massive stars.

5. Outlook

What is the origin of the magnetic fields in massive stars? It is quite possible that massive O-type stars behave similar to slowly rotating, chemically peculiar, magnetic Ap and Bp stars, which amount for 10 to 20% of main-sequence A and B stars. It was suggested that their magnetic fields are inherited from their parent molecular clouds (Moss 2003). Alternatively, from analysing possible paths for the formation of Ap and Bp stars based on modern theories for the evolution of single and binary stars, Tutukov & Fedorova (2010) suggested that merging of close binaries is their main formation channel.

Indeed, the evolution of a high fraction of massive stars is affected by environmental effects. Sana et al. (2012) find two thirds of all Galactic O stars to be members of close binary systems. Mass transfer or stellar merger may rejuvenate the mass gaining star while strong differential rotation may cause the creation of a magnetic field (Langer 2012). In dense star clusters, close multibody interactions may cause single and binary stars to be scattered out of the star cluster (e.g. Leonard & Duncan 1990).

Clearly, these various ideas bear significantly different predictions for the appearance of magnetic main sequence stars inside and outside of star clusters. This means that these predictions are well testable, and we are confident that the question of the origin of the magnetic fields in massive main sequence stars will be answered in the near future. As described in Sect. 4, first trends are emerging from the scarce data at hand. However, at the current stage, we clearly need more observations before firm conclusions can be drawn.


1

The spectropolarimetric capabilities of FORS 1 were moved to FORS 2 in 2009.

Acknowledgments

This research made use of the SIMBAD database, operated at CDS, Strasbourg, France. The Nordic Optical Telescope is operated on the island of La Palma jointly by Denmark, Finland, Iceland, Norway, and Sweden, in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias. We would like to thank G. Lo Curto for his help with the HARPS data reduction, and V. Gvaramadze for stimulating discussions. S.H. and J.F.G. acknowledge support by the Deutsche Forschungsgemeinschaft (Hu532/17-1).

References

All Tables

Table 1

List of massive stars observed for this programme.

Table 2

Longitudinal magnetic fields measured with FORS 2 in the studied sample.

Table 2

continued.

Table 3

Magnetic field measurements of five O-type stars using SOFIN and HARPS observations.

Table 4

Probable members in open clusters.

All Figures

thumbnail Fig. 1

Normalised FORS 1/2 Stokes I spectra of O-type stars with detected magnetic fields observed from 2005 to 2011. Well known spectral lines are indicated; all Balmer lines from the Balmer jump to Hβ are visible. The spectra were offset from 1 by multiples of 0.5 for clarity.

Open with DEXTER
In the text
thumbnail Fig. 2

Longitudinal magnetic field variation of the Of?p star HD 191612 according to the 537.6 d period determined by Howarth et al. (2007). Red symbols correspond to ESPaDOnS observations, while blue symbols are our SOFIN measurements (Hubrig et al. 2010, and Table 3). Note that the measurement errors for both ESPaDOnS and SOFIN observations are of similar order.

Open with DEXTER
In the text
thumbnail Fig. 3

Variable line profiles in the SOFIN spectra of HD 36879 obtained at two different epochs. Left panel: spectral variability of the Fe iii line profile at λ6048 and the blend Fe iii + He ii at λ6078. Right panel: spectral variability of line profiles presenting blends of the C iii and O ii lines at λ4647 and λ4660.

Open with DEXTER
In the text
thumbnail Fig. 4

Line profiles of He i in the spectra of HD 36879 obtained at two different epochs. From top left to bottom right we show the lines He i λ4713, λ4922, λ5015, and λ5876.

Open with DEXTER
In the text
thumbnail Fig. 5

Distinct variability observed in the spectra of CPD−28 2561 in the He i λ4026 (top left), He ii λ4686 (top right), Hβ (bottom left), and He i λ5876 (bottom right) lines.

Open with DEXTER
In the text
thumbnail Fig. 6

Longitudinal magnetic field variation of the Of?p star HD 148937 according to the 7.032 d period determined by Nazé et al. (2010). Red symbols correspond to the ESPaDOnS observations (Wade et al. 2012b), while green symbols are our FORS 1 and FORS 2 measurements (Hubrig et al. 2008; 2011b and present paper). Note that the measurement errors for both ESPaDOnS and FORS 1/2 observations are of similar order. One more measurement (not shown in this figure) was obtained in 2010 on May 22, but ordinary and extraordinary beams in the FORS 2 setup were overlapping in this exposure due to problems with slitlets.

Open with DEXTER
In the text
thumbnail Fig. 7

Phase diagram for the best sinusoidal fit corresponding to the period of 0.8 days for the longitudinal magnetic field measurements using hydrogen lines for ζ Oph.

Open with DEXTER
In the text
thumbnail Fig. 8

Phase diagram for the best sinusoidal fit corresponding to the period of 1.3 days for the longitudinal magnetic field measurements using the whole spectrum for ζ Oph.

Open with DEXTER
In the text
thumbnail Fig. 9

Combined IR Spitzer IRAC (3.6 μm blue, 4.5 μm green, 8.0 μm red) image of the bow shock around Vela X-1. Archival data have been used (AOR 17858304, PI: R. Iping). Image size is 5′ × 5′.

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.