arXiv:2010.00983v1 [astro-ph.SR] 2 Oct 2020 bu %o -yesaswt asseceig1 M 18 exceeding masses with probab O-type that of indicate 7% observations about im- path Previous understood, evolutionary fate. ’s well ultimate the not in remain uncertainties especial fields, large aspects, plying magnetic specific of stars, role massive the of modeling the In INTRODUCTION 1 Wolf-Rayet stars: — massive stars: — fields bu %o al -adOtp tr aemeasurable, have stars (e.g., O-type fields and magnetic B- dipolar early mostly of 6% about a erte elpae ihtertto eido 55hprevious h 15.5 of period rotation words: the Key with phased well rather be can © ihsrn antcfil eetosmyb eae omag- to related be with may netars detections field magnetic strong with h icvr favral antcfil nW 55 WR in field and magnetic stars Wolf-Rayet variable two a in of fields discovery magnetic the for search The MNRAS c¨ le ta.2017 al. Sch¨oller et oaiaino e ie 10 times few atmo- a star of WR indi- polarization in is present there of are that emi fields (e.g. X-ray fact spheres magnetic and variability the that spectral of sion, from e.g. spite evidence, rect in explored, ciently wind unobservable. almost dense surface their However, stellar the core. make now stellar and former envelope their evolved their expose chemically lost have highly that are stars o massive stars which predecessors direct WR stars, and WR remnants. stars in compact O fields massive of magnetic descendants detect are to important is it † eddtcina infiac ee f3 importance of exceptional level th significance of in a detection is at field environment, detection The molecular 55. field pr WR the in the Gauss and of hundred 78 search few RCW a the detect of for definite order 55, the formally WR of a and achieve 46 We WR observations. stars, tropolarimetric but WR presenc envelope, two the hydrogen selected on their present we expectations lost be the have substantiate should To stars unobservable. fields WR magnetic evolution, their that altho of emission, explored, well X-ray not and are stars variability (WR) Wolf-Rayet in fields Magnetic ABSTRACT ZZZ form original in YYY; Received XXX. Accepted 1 3 2 .Hubrig S. ebi-ntttfu srpyi osa AP,A e St der An (AIP), Potsdam f¨ur Astrophysik Leibniz-Institut hsc iiin ..Lwec eklyNtoa Laborat National Berkeley Lawrence E.O. Division, Physics Karl-Schwarzschild-Str. Observatory, Southern European orsodn uhr [email protected] author: Corresponding antcfilsi Rsasaecretyntsuffi- not currently are stars WR in fields Magnetic 09TeAuthors The 2019 aly&Ignace & Gayley 000 , B ihu ta.2014 al. et Michaux 1 1 † – ≈ 5 .Sch¨oller M. , ehius oaierc—sas niiul R4 tr:individual: stars: — 46 WR individual: stars: — polarimetric techniques: 21)Pern coe 00Cmie sn NA L MNRAS using Compiled 2020 October 5 Preprint (2019) 10 .Tertclmdl ugs htOstars O that suggest models Theoretical ). 15 ( 2010 (e.g. G rdce rcinlcircular fractional a predicted ) hmsne l 2004 al. et Thompson 2 .Peiu hoeia work theoretical Previous ). .Cikota A. , − 4 o antcfilsof fields magnetic for rnu ta.2017 al. et Grunhut σ a civdfrW 6 u h aiblt ftemaue edstre field measured the of variability the but 46, WR for achieved was 3 .P J¨arvinen P. S. , .Thus, ). ⊙ ,878Grhn,Germany Garching, 85748 2, rwre1,142Ptdm Germany Potsdam, 14482 16, ernwarte and and r,1CcornRa,Bree,C 42,USA 94720, CA Berkeley, Road, Cyclotron 1 ory, ly ly s- s f ; osac o ekmgei ed nanme fWR of number a in fields magnetic weak stars, for search to nor- continuum best the get to ne shapes is malization. order it adopt that to so orders, essary spe adjacent high-resolution over with extend observed tropolarimetry lines spectral broad The lal antcO? tr D183 n CPD and 148937 HD stars Of?p magnetic clearly eso pcrgahi pcrplrmti oe(OS2; (FORS mode spectropolarimetric in Spectrograph persion enlniuia antcfield magnetic longitudinal mean 1998 al. et Appenzeller s km thousand few a of by velocities lines wind emission with Th the wind, shifts of difficult. broadening stellar Doppler is wind stars strong the these is the problem in in major fields magnetic formed line of the is detection As stars the stars. of WR other limit in in upper G, non-detections 200 average spectrum the an for about and G 500 respectively, of about G, strengths 80 and 138, field WR G, 130 and magnetic 137, WR to 134, WR corresponding for detections field magnetic ylclyvral n -a mtigW5sa R6at 6 WR 3.3 star of WN5 level emitting significance X-ray a and variable cyclically evtoswt SaOSa h Canada–France–Hawaii the at ob Telescope, ESPaDOnS spectropolarimetric with high-resolution servations Using G. 100 about ercsetaallowed polari- FORS2 spectra obtained metric The Paranal/Chile. Cerro Telescope on Large Very (VLT) Obsevatory’s Southern European the of fmgei ed ntems-vle asv stars, massive most-evolved the in fields magnetic of e eas fsc mes iebodnn nW stars, WR in broadening line immense such of Because 1 urge al. et Hubrig o favral enlniuia antcfield magnetic longitudinal mean variable a of ion g hr sidrc vdne ..fo spectral from e.g. evidence, indirect is there ugh ysgetdb UEobservations. FUSE by suggested ly hi es id aeteselrcr almost core stellar the make winds dense their ssa,wihi soitdwt h ignebula ring the with associated is which star, is el hvoieee al. Chevroti´ere et la de sneo antcfilsuigFR spec- 2 FORS using fields magnetic of esence o u nesadn fsa omto.No formation. star of understanding our for nteesas en na dacdstage advanced an in Being stars. these in ( 2016 one nte8mAt telescope Antu m 8 the on mounted ) urge al. et Hubrig sdteFclRdcrlwdis- low Reducer FOcal the used ) σ epn nmn httetwo the that mind in Keeping . R5 tr:magnetic stars: — 55 WR h ( B 2014 z ( i 2016 258 = ,rpre marginal reported ), A omauethe measure to ) T E tl l v3.0 file style X ± 8Gi the in G 78 − 28 ◦ ngths 5104 − c- c- 1 e - . L2 Hubrig et al. were for the first time detected as magnetic in the FORS 2 of this star. Thus, the discovery of a magnetic field in WR 55 observations at significance levels of 3.1σ and 3.2σ, respec- would be of exceptional interest for star formation theories. tively (Hubrig et al. 2008, 2011), the detection of the mag- As WR stars are characterized by spectra showing very netic field in WR 6 at a significance level of 3.3σ indicates that broad emission lines, the determination of their magnetic a magnetic field is likely present in this star. Spectropolari- fields is usually based on the calculation of the mean lon- metric monitoring of WR 6 revealed a sinusoidal nature of the gitudinal magnetic field, i.e. of the line-of-sight field com- hBzi variations, which is indicative of a predominantly dipo- ponent, using circularly polarized light. To search for mag- lar magnetic field structure (Hubrig et al. 2016; see Fig. 1, left netic fields in WR 46 and WR 55, we obtained several ran- side). The field appeared to be reversing, with the extrema domly timed distributed low-resolution spectropolarimetric detected at rotation phases 0 and 0.5. observations using the FOcal Reducer low dispersion Spec- In the sample of spectropolarimetrically studied WR stars trograph (FORS 2; Appenzeller et al. 1998), installed at the by Hubrig et al. (2016), WR 6 was the only target showing ESO/VLT. In the following, we give an overview of our spec- X-ray emission and cyclical variability due to the presence tropolarimetric observations, describe the data reduction and of corotating interacting regions (CIRs), which are formed discuss the results of the magnetic field measurements. out of the interaction between high and low-velocity flows as the star rotates (e.g. St-Louis et al. 1995). CIRs were de- tected in spectroscopic time series observations of only a few 2 DATA REDUCTION AND RESULTS OF THE massive stars (e.g. Mullan 1984). It was suggested that CIRs MAGNETIC FIELD MEASUREMENTS are related to the presence of magnetic bright spots, which are indicators of the presence of a global magnetic field (e.g. To maximize the field detection probability and to avoid miss- Ramiaramanantsoa et al. 2014) ing the magnetic field due to an unfavorable viewing angle in certain rotation phases, FORS 2 observations of both targets To substantiate expectations on the presence of magnetic were obtained on several different epochs to sample differ- fields in the most-evolved massive stars, we have searched ent rotation phases. Seven spectropolarimetric observations for magnetic fields in two other promising targets, WR 46 of WR 46 were obtained in service mode, one observation on (=HD 104994) and WR 55 (=HD 117688), which, similar to 2016 March 14 and six observations from 2020 January 16 to WR 6, show CIRs (e.g. Chen´e& St-Louis 2011). These stars February 21. For WR 55, we obtained five observations from are accessible from the VLT and have never been observed 2020 February 13 to March 10. The last observation of WR 55 with spectropolarimetry in the past. recorded on March 10 was not completed, probably due to WR46 is a WN3p star (Hamann et al. 2019), with rela- bad weather conditions, and was therefore used only for the iv tively strong O λ 3811 and λ 3834 emission lines. It is very inspection of spectral variability. hot and compact (Teff = 112.2 kK, R = 1.4R⊙), and also The FORS 2 multi-mode instrument is equipped with po- bright in X-rays, possessing a hard component in its X-ray larisation analysing optics comprising super-achromatic half- emission (Gosset et al. 2011), which may indicate the mag- wave and quarter-wave phase retarder plates, and a Wollas- netic nature of WR 46. According to H´enault-Brunet et al. ton prism with a beam divergence of 22′′ in standard reso- (2011), WR 46 is known to exhibit a very complex variabil- lution mode. We used the GRISM 600B and the narrowest ity pattern. The different periods and timescales observed in available slit width of 0′′. 4 to obtain a spectral resolving power the past suggest the presence of multiple periods, including of R ∼ 2000 in the observed spectral range from 3250 to dominant and secondary periods (see Fig. 1 in their work). 6215 A.˚ For the observations, we used a non-standard read- To explain the short-term variability of this star, various sce- out mode with low gain (200kHz,1×1,low), which provides a narios were evoked, including the possibility of a close binary broader dynamic range, hence allowed us to reach a higher or non-radial pulsations (e.g. Veen et al. 2002a,b). Using ob- signal-to-noise ratio (S/N) in the individual spectra. The cir- servations with the Far Spectroscopic Explorer cular polarization observations were carried out using a se- (FUSE), H´enault-Brunet et al. (2011) found significant vari- quence of positions of the quarter-wave plate −45◦, +45◦, ations on a timescale of ∼8 h. This period is close to the +45◦, −45◦ and so forth, to minimize the cross-talk effect photometric and spectroscopic periods previously reported by and to cancel errors from different transmission properties other authors. H´enault-Brunet et al. (2011) also reported the of the two polarized beams. Moreover, the reversal of the detection of a second significant peak, just slightly weaker, quarter-wave plate compensates for fixed errors in the rela- corresponding to P = 15.5 ± 2.5 h. tive wavelength calibrations of the two polarized spectra. The WR 55 is a significantly cooler (Teff = 56.2kK) WN7 ordinary and extraordinary beams were extracted using stan- star with hydrogen deficiency and belongs to the WNE sub- dard IRAF procedures as described by Cikota et al. (2017). class, like WR 6 (Hamann et al. 2006). However, its radius, The wavelength calibration was carried out using He-Ne-Ar R = 5.2R⊙, is larger compared to the radius of WR 6 with arc lamp exposures. R = 3.2R⊙. A highly significant level of spectroscopic vari- The spectral appearance of the WN stars WR 46 and ability of about 10% was discovered by Chen´e& St-Louis WR 55 in the FORS 2 spectra is presented in Fig. 1. The (2011). So far, no periodicity search was carried out for spectra of WN stars are dominated by helium and nitrogen this star. Cappa et al. (2009) investigated the distribution lines. The WN3p star WR 46 is characterized by the presence of molecular gas related to the ring RCW 78 around of strong N v and He ii lines and the absence of hydrogen. The WR 55 and concluded that WR 55 is not only responsible for “p” stands for peculiar and denotes the presence of unusually the ionization of the gas in the nebula, but also for the cre- strong O vi λ 3811 and λ 3834 emission lines, the relatively ation of the interstellar bubble. Their analysis indicates that strong N v λ 4604 line, and relatively weak C iv λ 5801 and the star formation in this region is induced by the strong wind λ 5812 lines (e.g. Conti & Massey 1989).

MNRAS 000, 1–5 (2019) Magnetic fields in WR stars L3

3.0 2.5×105 2.0×105 2.5 1.5×105 2.0 C

ADUs 5

1.0×10 I/I 5.0×104 1.5 0 3500 4000 4500 5000 5500 6000 1.0 Wavelength [Å] 5850 5860 5870 5880 5890 5900 4×105 2.5 3×105

2×105 2.0 C ADUs I/I 5 1×10 1.5 0 3500 4000 4500 5000 5500 6000 1.0 Wavelength [Å] 4440 4460 4480 4500 4520 4540 4560 Figure 1. FORS 2 Stokes I spectra of WR 46 and WR 55. identification is based on the work of Hamann et al. (2006) and 3.0 is shown above the line profiles. 2.5 C

While previous observations of WR 46 clearly showed I/I 2.0 the presence of variability in and spectroscopy, the variability of WR 55 was studied only once by 1.5 Chen´e& St-Louis (2011) using spectra in the spectral range 5200–6000 A˚ obtained with the 1.5 m telescope of the Cerro 1.0 Tololo Inter-American Observatory. The authors reported for 4080 4100 4120 4140 this star a highly significant level of variability of up to 10% Wavelength [Å] of the line intensity. As we show in Fig. 2, spectral line vari- Figure 2. FORS 2 Stokes I spectra showing variability of different ability is also detected in our FORS 2 Stokes I spectra of this spectral lines. The spectra obtained at different epochs are offset star. vertically for better visibility. Spectral line identification is based A description of the assessment of the presence of a lon- on the work of Hamann et al. (2006). gitudinal magnetic field using FORS 1/2 spectropolarimet- ric observations was presented in our previous work (e.g. Hubrig et al. 2004a,b, and references therein). Improvements pose of deriving robust estimates of standard errors (e.g. to the methods used, including V/I spectral rectification and Steffen et al. 2014). The measurement uncertainties obtained clipping, were detailed by Hubrig et al. (2014). Null spectra before and after the Monte Carlo bootstrapping tests were are calculated as pairwise differences from all available V found to be in close agreement, indicating the absence of re- profiles so that the real polarisation signal should cancel out. duction flaws. The results of our magnetic field measurements From these, 3σ-outliers are identified and used to clip the V are presented in Table 1. profiles. This removes spurious signals, which mostly come For WR 46, the values for the longitudinal magnetic field from cosmic rays, and also reduces the noise. hBzi show change of polarity, with the strongest mean longi- The mean longitudinal magnetic field hBzi is measured on tudinal magnetic field of positive polarity hBzi = 342±154 G the rectified and clipped spectra, based on the relation fol- at a significance level of 2.2σ and the strongest field of neg- lowing the method suggested by Angel & Landstreet (1970): ative polarity hBzi = −199 ± 88 G at a significance level of 2.3σ. Since significant photometric and spectroscopic varia- 2 tions on a timescale of ∼8 h were reported in previous stud- V geff eλ 1 dI = − 2 hBzi , (1) ies of this star, assuming that this periodicity is caused by I 4πme c I dλ rotational modulation, we tested the distribution of the mea- where V is the Stokes parameter that measures the circular surement values over this period. We do not find any hint for polarization, I is the intensity in the unpolarized spectrum, sinusoidal modulation, which is expected for a large-scale or- geff is the effective Land´efactor, e is the electron charge, λ ganized dipole field structure. However, as we show in Fig. 3, is the wavelength, me is the electron mass, c is the speed rotation modulation is indicated in our data, if we use the of light, dI/dλ is the wavelength derivative of Stokes I, and period of 15.5 h suggested by H´enault-Brunet et al. (2011). hBzi is the mean longitudinal (line-of-sight) magnetic field. Due to the large uncertainty of this period, only the mea- Furthermore, we have carried out Monte Carlo bootstrap- surements obtained in 2020 are fitted by a sinusoid. The ping tests. These are most often applied with the pur- older measurement obtained in 2016 and marked by the filled

MNRAS 000, 1–5 (2019) L4 Hubrig et al.

Table 1. Longitudinal magnetic field values obtained for WR 46 hBzi = −378 ± 85 G at a significance level of 4.4σ and and WR 55 using FORS 2 observations. In the first column we show hBzi = 205 ± 58 G at a significance level of 3.5σ. In view the modified Julian dates of mid-exposures, followed by the corre- of the importance of the field detection at a significance level sponding signal-to-noise ratio (S/N) of the FORS 2 Stokes I spec- of 4.4σ, we decided to carry out a consistency check using a tra measured close to 4686 A.˚ The measurements of the mean lon- different spectral extraction. The parallel and perpendicular gitudinal magnetic field using the Monte Carlo bootstrapping test beams in the observations at this were extracted using and using the null spectra are presented in Columns 3 and 4. All a pipeline written in the MIDAS environment and developed quoted errors are 1σ uncertainties. by T. Szeifert, the very first FORS instrument scientist. More details on this pipeline can be found in Hubrig et al. (2014). MJD S/N hBzi hBzi N The result of this measurement, hBzi = −334 ± 77 G, is fully (G) (G) compatible with the measurement hBzi = −378±85 G within WR 46 the error bars. The simplest model for the magnetic field geometry in stars 57461.3213 2339 −199±88 46±81 with globally organized fields is based on the assumption that 58864.2960 1477 342±154 −68±146 58885.1674 1845 35±94 −62±109 the studied stars are oblique dipole rotators, i.e. their mag- 58885.2627 2046 −43±99 −7±97 netic field can be approximated by a dipole with the magnetic 58892.1191 1189 268±194 81±173 axis inclined to the rotation axis. Unfortunately, the rota- 58898.3843 1905 23±93 −41±115 tion axis inclination i for WR stars is undefined because of 58900.1569 1088 −112±160 23±179 their dense winds, making the measurement of the projected rotation velocity v sin i using broad emission lines impos- WR 55 sible. Since the rotation period and the limb-darkening are 58892.2058 2388 205±58 −16±55 also unknown for WR 55, we can only estimate for this star 58898.3492 2579 −378±85 −2±78 a minimum dipole strength of ∼1.13 kG using the relation 58900.2075 2168 56±80 63±82 Bd > 3 hBzimax (Babcock 1958). 58916.3059 2386 4±66 −16±57 58918.2531 533

3 DISCUSSION 600 Although magnetic fields are now believed to play an impor- WR 46 tant role in the evolution of massive stars, spectropolarimetric 400 observations of WR stars, which are descendants of massive O stars, are still very scarce. WR stars are usually rather faint, and, in addition, their line spectra are formed in the strong , with the wind broadening of the emission 200 lines up to a few thousand km s−1. Both WR 46 and WR 55

> [G] v >

z are faint with visual magnitudes m 10.9, and have never been observed spectropolarimetrically in the past. So far, the

MNRAS 000, 1–5 (2019) Magnetic fields in WR stars L5 lar to those found for the magnetic Of?p stars HD 148937 H´enault-Brunet V., St-Louis N., Marchenko S. V., Pollock and CPD −28◦ 2561 (Hubrig et al. 2008, 2011, 2013, 2015). A. M. T., Carpano S., Talavera A., 2011, ApJ, 735, 13 In these studies, not a single reported detection reached a 4σ Hubrig S., Kurtz D. W., Bagnulo S., Szeifert T., Sch¨oller M., significance level. Mathys G., Dziembowski W. A., 2004a, A&A, 415, 661 The first detection of the presence of a magnetic field Hubrig S., Szeifert T., Sch¨oller M., Mathys G., Kurtz D. W., 2004b, A&A, 415, 68 in WR 55 makes this star the best candidate for long-term Hubrig S., Sch¨oller M., Schnerr R. S., Gonz´alez J. F., Ignace R., spectropolarimetric monitoring. Future observing campaigns Henrichs H. F., 2008, A&A, 490, 793 should be based on spectropolarimetric time series to further Hubrig S., et al., 2011, A&A, 528, A151 strengthen the evidence for the magnetic nature of this star Hubrig S., et al., 2013, A&A, 551, A33 and to set more stringent limits to its magnetic field strength. Hubrig S., Sch¨oller M., Kholtygin A. F., 2014, MNRAS, 440, L6 The temporal variations of the measured longitudinal mag- Hubrig S., et al., 2015, MNRAS, 447, 1885 netic fields should be used to ascertain the rotation/magnetic Hubrig S., et al., 2016, MNRAS, 458, 3381 period of WR 55 and determine for the first time the geome- Michaux Y. J. L., Moffat A. F. J., Chen´eA.-N., St-Louis N., 2014, try of the global magnetic field in a WR star. MNRAS, 440, 2 Furthermore, the detection of the magnetic field in WR 55 Mullan D. J., 1984, ApJ, 283, 303 Ramiaramanantsoa T., et al., 2014, MNRAS, 441, 910 associated with the ring nebula RCW 78 and its molecular Sch¨oller M., et al., 2017, A&A, 599, A66 environment is of an exceptional importance for our under- St-Louis N., Dalton M. J., Marchenko S. V., Moffat A. F. J., Willis standing of star formation. According to Cappa et al. (2009), A. J., 1995, ApJ, 452, L57 WR 55 is not only responsible for the ionization of the gas in Steffen M., Hubrig S., Todt H., Sch¨oller M., Hamann W.-R., the nebula, but also for the creation of the interstellar bubble. Sandin C., Sch¨onberner D., 2014, A&A, 570, A88 The presence of star formation activity in the environment of Thompson T. A., Chang P., Quataert E., 2004, ApJ, 611, 380 this nebula suggests that it may have been triggered by the Veen P. M., Van Genderen A. M., Crowther P. A., van der Hucht expansion of the bubble. K. A. 2002a, A&A, 385, 600 Veen P. M., Van Genderen A. M., van der Hucht K. A. 2002b, A&A, 385, 619

ACKNOWLEDGEMENTS This paper has been typeset from a TEX/LATEX file prepared by the author. Based on observations made with ESO Telescopes at the La Silla Paranal Observatory under the programme IDs 097.D- 0428(A) and 0104.D-0246(A). SPJ is supported by the Ger- man Leibniz-Gemeinschaft, project number P67-2018. We thank the referee G. Mathys for his constructive comments.

DATA AVAILABILITY

The FORS 2 data from 2016 are available from the ESO Science Archive Facility at http://archive.eso.org/cms.html. The data from 2020 will become available in March 2021 at the same location and can be requested from the author be- fore that date.

REFERENCES

Angel J. R. P., Landstreet J. D., 1970, ApJ, 160, L147 Appenzeller I., et al., 1998, The ESO Messenger, 94, 1 Babcock H. W., 1958, ApJS, 3, 141 Cappa C. E., Rubio M., Mart´ın M. C., Romero G. A., 2009, A&A, 508, 759 Chen´eA.-N., St-Louis N., 2011, ApJ, 736, 140 Cikota A., Patat F., Cikota S., Faran T., 2017, MNRAS, 464, 4146 Conti P. S., Massey P., 1989, ApJ, 337, 251 de la Chevroti´ere A., St-Louis N., Moffat A. F. J., MiMeS Collab- oration, 2013, ApJ, 764, 171 de la Chevroti´ere A., St-Louis N., Moffat A. F. J., MiMeS Collab- oration, 2014, ApJ, 781, 73 Gayley K. G., Ignace R., 2010, ApJ, 708, 615 Gosset E., De Becker M., Naz´eY., Carpano S., Rauw G., Antokhin I. I., Vreux J.-M., Pollock A. M. T., 2011, A&A, 527, A66 Grunhut J. H., et al., 2017, MNRAS, 465, 243 Hamann W.-R., Gr¨afener G., Liermann A., 2006, A&A, 457, 1015 Hamann W.-R., et al., 2019, A&A, 625, A57

MNRAS 000, 1–5 (2019)