Systematic detection of magnetic fields in massive, late-type supergiants
Jason H. Grunhut, Gregg A. Wade, David A. Hanes, Evelyne Alecian
aa r X i v : . [ a s t r o - ph . S R ] J u l Mon. Not. R. Astron. Soc. , 1–8 (2009) Printed 10 October 2018 (MN L A TEX style file v2.2)
Systematic detection of magnetic fields in massive,late-type supergiants ⋆ J.H. Grunhut , † , G.A. Wade , D.A. Hanes , E. Alecian Department of Physics, Engineering Physics & Astronomy, Queens University, Kingston, Ontario, Canada, K7L 3N6 Department of Physics, Royal Military College of Canada, P.O. Box 17000, Station Forces, Kingston, Ontario, Canada, K7K 7B4 LAOG, Laboratoire d’Astrophysique de Grenoble, Universit´e Joseph Fourier, Grenoble Cedex, France
10 October 2018
ABSTRACT
We report the systematic detection of magnetic fields in massive (
M > M ⊙ ) late-type supergiants, using spectropolarimetric observations obtained with ESPaDOnSat the Canada-France-Hawaii Telescope. Our observations reveal detectable Stokes V Zeeman signatures in Least-Squares Deconvolved mean line profiles in one-third ofthe observed sample of more than 30 stars. The signatures are sometimes complex,revealing multiple reversals across the line. The corresponding longitudinal magneticfield is seldom detected, although our longitudinal field error bars are typically 0 . σ ). These characteristics suggest topologically complex magnetic fields, presumablygenerated by dynamo action. The Stokes V signatures of some targets show cleartime variability, indicating either rotational modulation or intrinsic evolution of themagnetic field. We also observe a weak correlation between the unsigned longitudinalmagnetic field and the Ca ii K core emission equivalent width of the active G2Iabsupergiant β Dra and the G8Ib supergiant ǫ Gem.
Key words: stars: supergiants - stars: magnetic fields - stars: late-type - techniques:spectropolarimetry
Supergiants are the descendants of massive O and B-type main sequence stars. When a massive star completesits main sequence evolution, it evolves rapidly across theHertzsprung-Russell (HR) diagram becoming a cool super-giant, characterized by a helium-burning core, and a deepconvective hydrogen-burning envelope. This is in strong con-trast to their OB main sequence phase, where they are char-acterized by a convective hydrogen-burning core and a radia-tive envelope (e.g Brown & Harper 2006). Subsequent evo-lution of the most massive stars (initial mass > M ⊙ ) in-volves the ejection of their envelope in the Wolf-Rayet phaseand a dramatic explosion in a type Ib supernova, while theless massive stars (initial mass between about 8 and 25 M ⊙ )evolve into red supergiants that explode as type II super-novae. Ultimately, the end product is a neutron star or blackhole (e.g. Crowther et al. 1995; Eldridge 2008).Due to their extended radii, low atmospheric densities, ⋆ Based on observations obtained at the Canada-France-HawaiiTelescope (CFHT) which is operated by the National ResearchCouncil of Canada, the Institut National des Sciences de l’Universof the Centre National de la Recherche Scientifique of France, andthe University of Hawaii. † E-mail: [email protected] slow rotation and long convective turnover times, super-giants provide an opportunity to study stellar magnetismand activity at the extremes of parameter space. Observa-tions of late-type supergiants (spectral type F and later)report an array of diverse and puzzling activity phenom-ena. Active late-type supergiants are luminous in X-raysand show emission in chromospheric ultraviolet (UV) Si iv lines, with some stars showing evidence of flaring - phenom-ena associated with the presence of a corona and magneticreconnection, presumably resulting from a dynamo-drivenmagnetic field (e.g. Tarasova et al. 2002; Ayres 2005). How-ever, unlike main sequence dwarfs, supergiants show little,if any, rotation-activity relation and are X-ray deficient byan order of magnitude or more compared to main sequencedwarf stars of similar spectral type (e.g. Ayres et al. 2005).There also exists a class of “non-coronal” or inactive super-giants that exhibit weak or no X-ray emission, UV spec-tra containing mostly low-temperature chromospheric lines(O i , Mg ii : T ∼ K) and much weaker UV fluxes inC iv chromospheric emission lines (Ayres et al. 2005). Theposition of these non-coronal stars on the HR diagram co-incides with the regions where cool (T × K) windsare predicted to dominate the circumstellar environment,potentially masking their X-ray emission (e.g. Haisch et al.1992; Ayres 2005). Another possible explanation for their c (cid:13) Grunhut et al. apparent lack of activity is that magnetic loops might besubmerged within the thick extended chromosphere (Ayres2005), resulting in internal absorption of coronal X-rays. Cu-riously, there also exist some “hybrid” stars that show thestrong C iv UV line emission of active supergiants, but thatare weak or undetected in X-rays (e.g Ayres 2005).Observations of the M2 supergiant Betelgeuse ( M =10 M ⊙ ; Carpenter & Robinson 1997) have revealed a chro-mosphere extending beyond 120 R ∗ , as well as evidenceof irregular surface brightness fluctuations, which simula-tions suggest are due to giant convective cells (Freytag et al.2002). Numerical simulations (Dorch 2004) predict thatthese convective regions are able to excite a dynamo capableof generating a highly structured surface magnetic field withlocalised features as strong as 500 G. In fact, Auri`ere et al.(2010) recently reported the detection of a magnetic field inthis star, with a (disc-integrated) longitudinal field compo-nent varying from about 0.5-1.5 G over about one month.Motivated by the outstanding puzzles associated withthe activity of late-type supergiants, the near complete lackof direct constraints on their magnetic fields, and recent suc-cesses measuring the magnetic fields of red and yellow giants(evolved intermediate-mass stars; e.g. Auri`ere et al. 2008,2009; Konstantinova-Antova et al. 2009), we have initiateda program to search for direct evidence of magnetic fields insupergiants (evolved high-mass stars). The data presented inthis paper represent stars with detected Zeeman signaturesobserved as part of a large survey of over 30 supergiantsranging in spectral type from early A- to late M. Circular polarisation (Stokes V ) spectra were obtained withthe ESPaDOnS spectropolarimeter, mounted on the 3.6mCanada-France-Hawaii Telescope (CFHT), as part of alarger survey investigating the magnetic properties of bright,late-type supergiants. In addition, one spectrum was ac-quired with the NARVAL spetropolarimeter on the BernardLyot Telescope (TBL).Both the ESPaDOnS and NARVAL instruments are op-tical echelle spectropolarimeters, capable of yielding broadbandpass (370 to 1050 nm), high resolution ( R = 68 000),polarised spectra. A complete polarisation observation con-sists of four individual sub-exposures between which thehalf-wave retarders are rotated back and forth between po-sition angles of -90 ◦ and +90 ◦ in order to reduce first-ordersystematic errors in the polarisation analysis. The extractionof the ESPaDOnS and NARVAL spectra, along with wave-length calibration and continuum normalisation, was ac-complished using the Upena pipeline running Libre-ESpRIT(Donati et al. 1997), a dedicated automatic reduction pack-age installed at both CFHT and TBL. To avoid saturation,we often obtained multiple consecutive sequences of obser-vations of a target that we co-added after extraction. Themedian signal-to-noise ratio (S/N) of the co-added obser-vations (computed from photon counting statistics, per 1.8km s − spectral pixel in the V -band) was about 2500, butranged as high as 4700.More than 30 stars have been observed as part of a sur-vey of the brightest supergiants. The current sample con-tains 4 A-type stars, 8 F-type stars, 11 G-type stars, 7 K- type stars, and 3 M-type stars. In this paper we report thecurrent results of this survey, with a focus on the 9 starsin which clear Stokes V Zeeman signatures have been de-tected. The journal of observations for the observed stars isprovided in Table 1.
In order to increase the S/N and detect weak Zeeman sig-natures in the circularly polarized spectra, we applied theLeast-Squares Deconvolution (LSD; Donati et al. 1997) pro-cedure to all polarimetric observations. The line masks usedin this analysis were constructed for photospheric lines usingatomic data from the Vienna Atomic Line Database (VALD;Ryabchikova et al. 1997; Kupka et al. 1999) with line depthscomputed assuming solar abundances and LTE. Only lineswith predicted depths greater than 10 percent of the con-tinuum flux were included in the line masks. This resultedin, for example, about 4000 lines included in the mask for aG2 star. The final result of the LSD analysis of a sequenceof observations is a single, mean circular polarisation pro-file (LSD Stokes V ), a mean diagnostic null profile (LSD N ), and a mean un-polarized profile (LSD Stokes I ). Thisresulted in an increase in the S/N by a factor of 21 to 68,roughly varying as the square root of the number of lines inthe mask. All LSD profiles were produced on a spectral gridwith a velocity bin of 1.8 km s − .Fig. 1 illustrates the LSD profiles of 9 stars of our sam-ple for which magnetic fields were unambiguously detected,showing the often complex Zeeman signatures across thespectral lines of supergiants from early-F to mid-K spec-tral type and their placement on the HR diagram. Tem-peratures and luminosities were compiled from various pub-lished catalogues (Cayrel de Strobel et al. 2001; Gray et al.2001; Kovtyukh 2007; Lyubimkov et al. 2010), or otherwisederived using UBV photometry from the Bright Star Cat-alogue (Hoffleit & Warren 1991) using transformations ofFlower (1996) and Bessell et al. (1998). Six additional starsin which no statistically significant Zeeman detection isachieved but in which the presence of magnetic field is sus-pected (based on an excess of Stokes V signal at the positionof the stellar mean spectral line) are shown in Fig. 2. Finally,the remaining stars presenting no evidence of a Zeeman sig-nature are illustrated in Fig. 3.The longitudinal field measurements reported in Table 1were measured from the LSD Stokes I and V profiles in themanner described by Silvester et al. (2009), with the inte-gration carried out within the limits of the Stokes I lineprofile. We stress that while the longitudinal field providesa useful measure of the line-of-sight component of the field,we do not use it as the primary diagnostic of the presence ofa magnetic field. This is because a large variety of magneticconfigurations can produce a null longitudinal field (as isapparent from our measurements in Table 1), while many ofthese configurations will generate a detectable Stokes V sig-nature in the velocity-resolved line profiles that we observed,as is evident in Fig. 1.We find an increase in the number of detectableStokes V signatures for observations of higher S/N (the peakS/N of the unpolarized spectra in the V -band ranges from700 to 3300). We find only 4 detections ( ∼
40 percent of our c (cid:13) , 1–8 ystematic detection of magnetic fields in massive, late-type supergiants Figure 1.
Hertzsprung-Russell diagram (center frame) showing all supergiants observed as part of our survey for which parameters wereavailable, with evolutionary tracks of Schaller et al. (1992). Black squares indicate stars for which no clear Zeeman signature is found,grey triangles indicate stars for which a Zeeman signature is suspected, and grey stars indicate stars with a clear Zeeman signature.Surrounding the HR diagram are illustrative LSD Stokes I (bottom), Stokes V (top) and diagnostic null (middle) profiles for the 9 starswith unambiguously detected Stokes V signatures.c (cid:13) , 1–8 Grunhut et al.
Table 1.
Journal of ESPaDOnS/NARVAL observations. Columns 1-7 list the target name, the HD number, the spectral type (fromSIMBAD), the observation date, the exposure time, the peak signal-to-noise ratio (S/N) in the V -band in the observed spectrum, and themean S/N ratio (per 1.8 km s − velocity bin) in the LSD Stokes V profile, for each observation. Columns 8-10 list the mean longitudinalfield measurement from the LSD profiles, the false alarm probability (FAP), and the Stokes V detection diagnosis (DD=Definite Detection,MD=Marginal Detection, ND=No Detection) as described by Donati et al. (1997). The first 9 rows represent stars in which we observeclear Zeeman signatures in Stokes V , while the next 6 rows represent stars with suspected Zeeman signatures. The last rows representstars with no visible Zeeman signature in Stokes V .Name HD Spec. Obs. Exp. Time S/N S/N h B z i ± σ B FAP Det.Type Date (s) V LSD V (G) Flag α Lep 36673 F0Ib 2009-09-27 3 × ×
32 2709 65870 0 . ± .
37 2.420E-07 DD α Per 20902 F5Iab 2010-01-26 8 × ×
15 4661 130558 0 . ± . < η Aql 187929 F6Iab 2009-09-05 3 × ×
111 2488 61608 − . ± .
75 6.985E-08 DD β Dra 159181 G2Iab 2010-03-05 6 × ×
36 4526 131246 − . ± . < ξ Pup 63700 G6Ia 2010-01-26 4 × ×
61 3543 112130 0 . ± . < ǫ Gem 48329 G8Ib 2010-01-25 5 × ×
43 4262 130843 − . ± . < c Pup 63032 K2.5Ib-II 2009-11-28 3 × ×
55 2188 65948 1 . ± .
39 1.410E-06 DD32 Cyg 192909 K3Ib+ 2009-10-02 2 × ×
103 1530 50481 1.16 ± λ Vel 78647 K4.5Ib-II 2009-05-07 10 × × . ± .
33 5.995E-15 DD β Aqr 204867 G0Ib 2009-09-06 3 × ×
164 2673 80881 0 . ± .
31 1.138E-02 ND α Aqr 209750 G2Ib 2009-09-06 3 × ×
43 3313 77144 0.47 ± d Cen 117440 G9Ib 2010-06-04 3 × ×
73 2464 76850 0.27 ± ξ Cyg 200905 K4.5Ib-II 2009-10-02 2 × ×
80 1640 54037 − . ± .
36 2.850E-02 ND α Ori 39801 M2Iab 2009-09-28 × × ± σ CMa 52877 M1.5Iab 2009-10-02 2 × ×
49 2033 66480 0.61 ± η Leo 87737 A0Ib 2009-12-07 3 × ×
93 3069 48084 0.03 ± α Cyg 197345 A2Iae 2009-09-05 3 × ×
44 2861 60421 2.47 ± L Pup 62623 A3Iab 2009-12-09 4 × ×
104 2749 55470 -9.76 ± ǫ Aur 31964 A8Iab 2009-09-28 3 × ×
47 2671 56858 1.27 ± LN Hya 112374 F3Ia 2009-05-06 1 × ×
491 664 12696 -2.98 ± B Vel 74180 F3Ia 2010-03-08 3 × ×
80 1789 33932 1.47 ± δ Cep 213306 F5Iab 2009-10-02 3 × ×
97 1726 37934 -0.25 ± γ Cyg 194093 F8Ib 2009-09-06 3 × ×
23 2738 75220 -0.24 ± δ CMa 54605 F8Iab 2009-09-27 3 × ×
64 2754 74751 0.37 ± × ×
185 1448 42557 -0.70 ± ζ Gem 52973 G0Ibv 2009-09-10 3 × ×
100 2346 54880 0.29 ± F Hya 74395 G1Ib 2010-01-28 2 × ×
201 2553 75160 0.29 ± β Pyx 74006 G7Ib-II 2009-11-24 3 × ×
88 2706 89445 0.93 ± × ×
72 1878 60586 -0.05 ± × ×
320 681 21839 -2.73 ± ǫ Peg 206778 K2Ib 2008-12-20 5 × ×
15 479 30740 0.03 ± π Pup 56855 K3Ib 2009-11-28 2 × ×
23 1641 52060 -1.24 ± α Sco 148478 M1.5Iab-b 2010-01-25 2 × × ± The data presented in this paper are the co-addition of observations taken on 2009-09-28, 2009-10-02, and 2009-10-07. detected number of stars) with peak S/N of 2500 or less (thebottom half of our sample), while we find 5 detections ( ∼ α Per with a peakS/N of 1509 in the V -band. Nevertheless, based on our ownexperience and that of Auri`ere et al. (2010) it appears thatachieving a S/N of 2500 or higher is best suited to detectingand characterising magnetic fields in supergiants using themethods described here. Our investigation of late-type supergiants shows that manyhost detectable Stokes V Zeeman signatures. The signa-tures are frequently complex, and the associated longitu- dinal magnetic field are generally weaker than 1 G. We havedetected usually unambiguous Zeeman signatures in 9 starsof our sample, with 6 additional stars suspected to showsignatures. The detected stars span a large range of physi-cal characteristics, with the most massive star detected be-ing α Lep (F0Ib, ∼ M ⊙ ). This star also represents thehottest star in our sample with a detectable circular po-larisation Zeeman signature ( T eff ∼ β Dra (G2Iab, ∼ M ⊙ ). Based on our adopted effec-tive temperatures, the coolest star with a detection is c Pup( T eff ∼ λ Vel.As is evident from the HR diagram in Fig. 1, we ob-tain a high incidence fraction for stars with spectral typesF ( ∼
40 percent; total of 8 stars observed), G ( ∼
30 percent;total of 11 stars observed), and K ( ∼
40 percent; total of 7 c (cid:13) , 1–8 ystematic detection of magnetic fields in massive, late-type supergiants Figure 2.
LSD Stokes I (bottom), Stokes V (top) and diagnostic null (middle) profiles for stars with suspected Zeeman signaturesdetected in Stokes V (based on an excess of Stokes V signal at the position of the stellar mean spectral line), which still result innon-detections based on the detection criteria of (Donati et al. 1997). stars observed). For M-type supergiants this fraction is moreuncertain (with a total of just 3 stars observed); with twosuspected detections the incidence fraction may be as high67 percent. We currently have not obtained any detectionsin hotter A-type stars (total of 4 stars observed).Prior to November 2009, ESPaDOnS was known to suf-fer from (variable) polarisation cross-talk at a level of 1-4percent. During the period when most of the observationsdescribed here were acquired, the primary source of thiscross-talk was the ESPaDOnS atmospheric dispersion cor-rector (ADC; Barrick et al. 2010). In November 2009, a newADC was installed in front of the ESPaDOnS polarimetricmodule, and the measured cross-talk since that time has re-mained stable at a level of about 0.6 percent. The repeatabil-ity of our Zeeman detections (both with ESPaDOnS, beforeand after resolution of the cross-talk issue, as well as withNarval), the existence of both detections and non-detectionswithin our sample, the diversity and complexity of the de-tected signatures, and the lack of any published evidencefor strong, coherent linear polarisation within the metallicabsorption lines of late-type supergiants, makes us confidentthat the Stokes V signatures detected within this survey arenot significantly affected by cross-talk.Our current dataset reveals no clear differences betweenthe classical optical activity indicators (such as Ca ii H&Kemission or H α emission) of those stars for which we detectfields and those for which we do not. Additionally, we find no obvious differences between the rotational velocities ofthe stars with magnetic detections and those without (using v sin i measurements of de Medeiros et al. (2002)). However,there does appear to be a weak correlation between the un-signed longitudinal field strength and Ca ii core equivalentwidth measurements for some stars in which fields are de-tected on multiple dates, such as β Dra and ǫ Gem (thiscorrelation for β Dra is shown in Fig. 4).Historically, a number of searches for magnetic fields inlate-type supergiants have been published (e.g. Borra et al.1981; Scholz & Gerth 1981; Plachinda 2005). Essentially,our direct detections of magnetic fields in supergiants - inparticular our typical observation of sub-1 G longitudinalfields - are not consistent with any of the reports claimed inthose papers.A few of the stars in our sample were previously investi-gated for magnetic fields. Several previous magnetic analysesof α Per were unable to detect a significant field, with thelowest uncertainty that of Shorlin et al. (2002) with h B z i =1 ± B z = 0 ±
49 G). Our 5 observations of this star spanalmost a full year, show some variability (see Fig. 5) andconfirm the presence of a weak longitudinal field ( ∼ V profile of all observed stars, with anamplitude larger by a factor of ∼ η Aql was carried out by c (cid:13) , 1–8 Grunhut et al.
Figure 3.
LSD Stokes I (bottom), Stokes V (top) and diagnostic null (middle) profiles for stars with no evidence of a Zeeman signaturein Stokes V . c (cid:13) , 1–8 ystematic detection of magnetic fields in massive, late-type supergiants Figure 3 – continued LSD Stokes I (bottom), Stokes V (top)and diagnostic null (middle) profiles for stars with no evidence ofa Zeeman signature in Stokes V . -0.5 0 0.5 1 1.5 2 2.5 3 3.5 Unsigned B z (G) C a II E qu i v a l e n t W i d t h Figure 4. Ca ii K line core equivalent width measurements for β Dra as a function of the unsigned longitudinal magnetic field(B z ). Wade et al. (2002) to investigate the reported detections byPlachinda (2000). Wade et al. found an average unsignedlongitudinal field strength of 10 G or less and a typical un-certainty of 5 G, while Plachinda claimed maximum fieldstrengths of ∼
100 G, varying with the pulsation period. Ourobservation of this star does show a clear signature (seeFig. 1), corresponding to a null longitudinal field measure-ment ( B z = − . ± .
75 G) consistent with the low valuesof Wade et al. near the same pulsation phase (phase 0.76).We note the possibility that the velocity field due to pul-sation is responsible for the unusual shape of the observedStokes V profile. ǫ Gem was investigated by Plachinda (2005), who re-ported the detection of a variable field with unsigned fieldstrengths ranging from ∼ ∼
50 G, with a typical uncer-tainty ∼ V , our measured longitudinal fields of this star aresignificantly lower than the detections claimed by Plachinda(2005). The observed Stokes V profile of ǫ Gem appears tovary on the order of months, as illustrated in Fig. 6.Several supergiants for which we have detected a Zee-man signature are also known pulsators or variable stars(such as α Lep (Arkharov et al. 2005), η Aql (Kiss & Vink´o2000), ξ Pup (Koen & Eyer 2002)). Curiously, α Per is anon-pulsating star that clearly lies within the instability -200 -100 0 100 velocity (km/s) V / I c α Per (F5Iab) -200 -100 0 100 200 velocity (km/s) I / I c Figure 5.
Stokes V profiles (left, solid lines) and Stokes I (right)of α Per for the nights indicated. Profiles are vertically offset fordisplay purposes. The dashed line corresponds to the observationobtained on 04 Dec. 2009, shifted to the position of each night inorder to highlight changes in the profile. -200 -100 0 100 velocity (km/s) -5e-0505e-050.00010.000150.00020.00025 V / I c ε Gem (G8Ib) -200 -100 0 100 200 velocity (km/s) I / I c Figure 6.
Stokes V profiles (left, solid lines) and Stokes I (right)of ǫ Gem for the nights indicated. Profiles are vertically offset fordisplay purposes. The dashed line corresponds to the observationobtained on 15 Feb. 2009, shifted to the position of each night inorder to highlight changes in the profile. strip (Keller 2008) with evidence of circumstellar shells (e.g.Neilson & Lester 2008) and a stronger Zeeman signature.A few of the stars in our sample fall into the “hy-brid” chromosphere category, such as α Aqr, β Aqr, and c Pup. All of these stars have weak but detectable X-ray emission (Ayres 2005). While neither α Aqr or β Aqrhave detected Zeeman signatures in Stokes V , our obser-vation do show complex Stokes V profiles that are sugges-tive of a magnetic field. Comparing with the active stars β Dra or 32 Cyg (for which UV prominences were detectedby Schroeder (1983)), we see no clear differences in thestrengths of their Stokes V signatures that would reflectdifferences in their magnetic fields. We also observe a clearsignature in λ Vel, a “non-coronal” star with a typical coolchromosphere (Carpenter et al. 1999). c (cid:13) , 1–8 Grunhut et al.
Our survey of magnetic fields in massive, late-type super-giants reveals detectable Zeeman Stokes V signatures in theLSD profiles of about one-third of our sample of more than30 stars. The signatures are sometimes complex and corre-spond to longitudinal magnetic fields generally below 1 G.These characteristics suggest topologically complex mag-netic fields, presumably generated by dynamo action. Giventhe high rate of incidence, it may well be that magneticfields are excited in all cool supergiants. Nevertheless, ourfailure to detect fields in some stars for which very high S/Nobservations were acquired points to a large range of fieldstrengths or complexities. In fact, Auri`ere et al. (2010) re-port the detection of a clear Zeeman signature in Stokes V spectra of the cool M2 supergiant Betelgeuse, with a S/Nalmost twice has high as typically achieved in this study.This leads us to believe that a S/N approximately twice ashigh as that achieved here would be valuable for a morecomplete assessment of field incidence. An additional inter-esting result of this study is the detection of Zeeman sig-natures in hybrid and non-coronal supergiants - signatureswith strength and structure similar to those observed in ac-tive supergiants. As similar magnetic fields are therefore in-ferred to exist in all three classes of stars, this suggests thatthe magnetic properties of these different classes are similarand that other phenomena (e.g. attenuation by cool windsor extended chromospheres) are likely the cause of the ob-served differences in their activity. ACKNOWLEDGMENTS
The authors thank Dr. John Landstreet for helpful dis-cussion and effective motivation, and the referee Dr. TonyMoffat for constructive advice. JHG acknowledges finan-cial support in the form of an Ontario Graduate Scholar-ship. GAW and DAH acknowledge Discovery Grant supportfrom the Natural Science and Engineering Research Coun-cil of Canada. This research has made use of the SIMBADdatabase, operated at CDS, Strasbourg, France.
REFERENCES
Arkharov A.A., Hagen-Thorn E.I., Ruban E.V., 2005,ARep, 49, 526Auri`ere M., Konstantinova-Antova R., Petit P., et al., 2008,A&A, 491, 499Auri`ere M., Wade G.A., Konstantinova-Antova R., et al.,2009, A&A, 504, 231Auri`ere M., Donati J.-F., Konstantinova-Antova R., PerrinG., Petit P., Roudier T., 2010, A&A, 516, L2Ayres T.R., 2005, ApJ, 618, 493Ayres T.R., Brown A., Harper G.M., 2005, ApJ, 627, L53Barrick G., Benedict T., Sabin D., Manset N., Wade G.A.,2010, PASP, submittedBessell M.S., Castelli F., Plez B., 1998, A&A, 333, 231Borra E.F., Landstreet J.D., 1973, ApJ, 185, L139Borra E.F., Fletcher J.M., Poeckert R., 1981, ApJ, 247, 569Brown A., Harper G.M., 2006, in Sonneborn G., Moos H.,Andersson B.-G. (eds), ASP Conf. Ser. 345, Astrophysicsin the Far Ultraviolet. p. 165 Carpenter K.G., Robinson R.D., 1997, ApJ, 479, 930Carpenter K.G., Robinson R.D., Harper G.M., BennetP.D., Brown A., Mullan D.J., 1999, ApJ, 521, 382Cayrel de Strobel G., Soubiran C., Ralite N., 2001, A&A,373, 159Crowther P.A., Hillier D.J., Smith L.J., 1995, A&A, 293,172de Medeiros J.R., Udry S., Burki G., Mayor M., 2002,A&A, 395, 97Donati J.-F., Semel M., Carter B.D., Rees D.E., CollierCameron A., 1997, MNRAS, 291, 658Dorch S.B.F., 2004, A&A, 423, 1101Eldridge J.J, 2008, RSPTA, 366, 4441Flower P.J., 1996, ApJ, 469, 355Freytag B., Steff M., Dorch B., 2002, Astron. Nachr., 323,213Gray R.O., Graham P.W., Hoyt S.R., 2001, AJ, 121, 2159Haisch B., Schmitt J.H.M.M, Fabian A.C, 1992, Nature,360, L239Hoffleit D., Warren Jr W.H., 1991, Astronomical Data Cen-terKeller S.C., 2008, ApJ, 677, 483Kiss L.L, Vink´o J., 2000, MNRAS, 314, 420Koen C., Eyer L., 2002, MNRAS, 331, 45Konstantinova-Antova R., Auri`ere M., Schroder K.P., PetitP., 2009, in Strassmeier K.G., Kosovichev A.G., BeckmanJ. (eds), Proc. IAU Symp. 259, Cosmic Magnetic Fields.Cambridge Univ. Press, Cambridge, p. 433Kovtyukh V.V., 2007, MNRAS, 378, 617Kupka F., Piskunov N.E., Ryabchikova T.A., StempelsH.C., Weiss W.W., 1999, A&AS, 138, 119Lyubimkov L.S., Lamber D.L., Rostopchin S.I.,Rachkovskaya T.M., Pokland D.B., 2010, MNRAS,402, 1369Neilson H.R., Lester J.B., 2008, ApJ, 684, 569Plachinda S.I., 2000, A&A, 360, 642Plachinda S.I., 2005, Ap, 48, 9Ryabchikova T.A., Piskunov N.E., Kupka F., Weiss W.W.,1997, Baltic Astronomy, 6, 244Schaller G., Schaerer D., Meynet G., Maeder A., 1992,A&AS, 96, 269Scholz G., Gerth E., 1981, MNRAS, 195, 853Schroeder K.P., 1983, A&A, 124, L16Shorlin S.L.S., Wade G.A., Donati J.-F., Landstreet J.D.,Petit P., Sigut T.A.A., Strasser S., 2002, A&A, 392, 637Silvester J., Neiner C., Henrichs H.F., et al., 2009, MNRAS,398, 1505Tarasova T.N., 2002, ARep, 46, 474Wade G.A., Chadid M., Shorlin S.L.S., Bagnulo S., WeissW.W., 2002, A&A, 392, L17 c (cid:13)000