MOBSTER: Identifying Candidate Magnetic O Stars through Rotational Modulation of TESS Photometry
James Barron, Gregg A. Wade, Dominic M. Bowman, Alexandre David-Uraz, Melissa S. Munoz, Herbert Pablo, Sergio Simón-Díaz
MMOBSTER: Identifying CandidateMagnetic O Stars through RotationalModulation of
TESS
Photometry
James Barron · , Gregg A. Wade , Dominic M. Bowman , Alexandre David-Uraz ,Melissa S. Munoz , Herbert Pablo and Sergio Sim´on-D´ıaz ·
1. Department of Physics, Engineering Physics & Astronomy, Queen’s University, 64 Bader Lane, Kingston, ON, K7L 3N6, Canada2. Department of Physics and Space Science, Royal Military College of Canada, PO Box 17000, Kingston, ON, K7K 7B4, Canada3. Institute of Astronomy, KU Leuven, Celestijnenlaan 200D, B-3001 Leuven, Belgium4. Department of Physics and Astronomy, University of Delaware, Newark, DE 19716, USA5. AAVSO, 49 Bay State Road, Cambridge, MA 02138, USA6. Instituto de Astrof´ısica de Canarias,E-38 200 La Laguna, Tenerife, Spain7. Universidad de La Laguna, Universidad de La Laguna, E-38 205 La Laguna, Tenerife, Spain
Being relatively rare, the properties of magnetic O stars are not fully understood.To date fewer than a dozen of these stars have been confirmed, making any in-ference of their global properties uncertain due to small number statistics. Tobetter understand these objects it is necessary to increase the known sample. TheMOBSTER collaboration aims to do this by identifying candidate magnetic O,B, and A stars from the identification of rotational modulation in high-precisionphotometry from the Transiting Exoplanet Survey Satellite (
TESS ). Here we dis-cuss the collaboration’s efforts to detect rotational modulation in
TESS targets toidentify candidate magnetic O stars for future spectropolarimetric observations.
O-type stars are among the most massive and luminous stars with surface temper-atures over 30 000 K. They are the progenitors of black holes and neutron stars andprovide chemical enrichment to the surrounding interstellar medium. Unlike lowmass stars that generate their magnetic fields through convective dynamos, high-mass stars possess radiative envelopes and lack convection in their outer layers togenerate large scale magnetic fields. Nevertheless, spectropolarimetric observationshave shown that some O stars possess strong ( > fossil fields (Borra et al., 1982).The origins of the fossil fields are still debated, whether they are formed duringpre-main sequence evolution (Villebrun et al., 2019) or in stellar mergers (Schneideret al., 2019). These surface magnetic fields affect a star’s evolution through bothmagnetic braking (Ud-Doula et al., 2009) and interactions with the stellar windleading to the formation of a magnetosphere (ud-Doula & Owocki, 2002).To date there are only 11 confirmed magnetic O stars (Wade & MiMeS Collabora-tion, 2015), giving a magnetic incidence rate of less than 10% (Grunhut et al., 2017).The small number of stars available for study makes it difficult to model their evolu-tion and infer global properties. It is necessary to identify more magnetic O stars toincrease our knowledge about their origin and evolutionary paths. The Transiting pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63) January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
O-type stars are among the most massive and luminous stars with surface temper-atures over 30 000 K. They are the progenitors of black holes and neutron stars andprovide chemical enrichment to the surrounding interstellar medium. Unlike lowmass stars that generate their magnetic fields through convective dynamos, high-mass stars possess radiative envelopes and lack convection in their outer layers togenerate large scale magnetic fields. Nevertheless, spectropolarimetric observationshave shown that some O stars possess strong ( > fossil fields (Borra et al., 1982).The origins of the fossil fields are still debated, whether they are formed duringpre-main sequence evolution (Villebrun et al., 2019) or in stellar mergers (Schneideret al., 2019). These surface magnetic fields affect a star’s evolution through bothmagnetic braking (Ud-Doula et al., 2009) and interactions with the stellar windleading to the formation of a magnetosphere (ud-Doula & Owocki, 2002).To date there are only 11 confirmed magnetic O stars (Wade & MiMeS Collabora-tion, 2015), giving a magnetic incidence rate of less than 10% (Grunhut et al., 2017).The small number of stars available for study makes it difficult to model their evolu-tion and infer global properties. It is necessary to identify more magnetic O stars toincrease our knowledge about their origin and evolutionary paths. The Transiting pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63) January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63) a r X i v : . [ a s t r o - ph . S R ] J a n ames Barron, et al.Exoplanet Survey Satellite ( TESS ; Ricker et al., 2015) offers an unparalleled oppor-tunity to study stellar variability across the HR diagram due to its precision, highcadence and comprehensive coverage of the sky ( ∼ TESS : probing their Evolutionary and Rotationalproperties; David-Uraz et al., 2019b; Sikora et al., 2019; Shultz et al., 2019) aims tomake use of
TESS data to search for rotational modulation in OBA stars to iden-tify magnetic massive star candidates for follow-up spectropolarimetric observations.For further discussion about the collaboration and rotational modulation as seen inB and A-type stars see David-Uraz et al. (2019a) and David-Uraz et al. (2019c).Here we discuss our efforts towards identifying magnetic O star candidates throughrotational modulation in
TESS photometry.
In OB stars that possess surface magnetic fields, the radiatively-driven stellar windis confined by the closed field lines around the magnetic equator, forming a mag-netosphere. As the star rotates, the column density along the line-of-sight changesdue to the misalignment between the rotational and magnetic axes. Therefore thecontinuum intensity of light varies during the rotation phase, leading to periodicvariations in the star’s photometric light curve (e.g. Munoz et al., 2019). Rotationalmodulation is also seen in later B and A type stars. In this case the magnetic fieldsinfluence the atomic diffusion process in the star’s atmosphere (Alecian & Stift,2010), leading to chemical inhomegeneities on the stellar surface which are seen asbrightness spots. Such variations have been successfully detected in known magneticB stars (David-Uraz et al., 2019b) using a Lomb-Scargle (L-S) analysis (see Vander-Plas 2018 for an introduction to the technique). The rotational signature typicallymanifests itself in the periodogram as a peak at the rotation frequency and a numberof harmonics, which depend on the spatial distribution of spots on the stellar surfacewith respect to the observer (e.g. Stibbs, 1950; Bowman et al., 2018; Sikora et al.,2019).However, such a signature is not unique to rotational modulation, and can alsobe associated with ellipsoidal variations in binary systems. Before
TESS , high pre-cision space photometric surveys of O stars, while limited, have found diverse typesof variability including periodic variations, either due to rotation or binarity andstochastic low-frequency signals (e.g. Blomme et al., 2011; Buysschaert et al., 2015).These stochastic signals have been attributed to both internal gravity waves (e.g.Aerts & Rogers, 2015; Bowman et al., 2019a,b) and subsurface convection zones(Lecoanet et al., 2019). A study of the variability of O and B stars using 2-min
TESS data for sectors 1 and 2 has been conducted by Pedersen et al. (2019) andcontained 5 O star targets, of which 3 were found to exhibit rotational modulation.To aid in the diagnosis of rotational modulation, other criteria must be consid-ered. Literature searches and examination of available spectra can help diagnosevariability due to stellar companions. The projected rotational velocity found fromspectroscopy provides a lower limit for the star’s rotational frequency, and an upperlimit can be placed by the star’s critical rotation velocity. Light curves that are foundto show rotational modulation can then be modelled to constrain parameters relatingto the star’s magnetic field to help determine the feasibility of spectropolarimetricfollow-up observations. The Analytic Dynamical Magnetosphere (ADM) model de-52 (cid:63)
PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 pta.edu.pl/proc/2020jan15/123
OBSTER: Identifying Candidate Magnetic O Stars through . . .veloped by Owocki et al. (2016) offers a time-averaged description of the density andvelocity structure of the magnetosphere of slowly rotating magnetic massive stars.The model is in good agreement with magnetohydrodynamic simulations and can beused to determine magnetic and stellar parameters from photometry (Munoz et al.,2019). The existence of rotational modulation alone does not necessarily imply theexistence of a magnetosphere, but it can be used as a diagnostic tool to identifycandidate magnetic O-type stars. TESS
Observations
The
TESS survey divides the sky into the north and south ecliptic hemispheres,each containing 13 partially overlapping sectors that are observed nearly contin-uously for 27 days. Target pixel files (TPFs) for 200 000 pre-selected targets areavailable at 2-min cadence, and full-frame images (FFIs) are available at a 30-mincadence for approximately 470 million point sources (Stassun et al., 2018). In addi-tion to the TPFs, 2-min processed (
PDCSAP ) light curves are provided by the
TESS
Science Team (Jenkins et al., 2016) through the Mikulski Archive for Space Tele-scopes (MAST). It is important to note that
TESS pixels are relatively large (21 ×
21 arcsec), and so targets may include flux from multiple nearby stars (e.g. Fig1). Light curve extraction from
TESS
FFIs can be done using open-source toolssuch as eleanor (Feinstein et al., 2019). The tool can perform background sub-traction, instrument systematics decorrelation, principal component analysis, andpoint-spread function modelling. As of July of 2019 observations of the southernecliptic were concluded, allowing for the opportunity to perform a comprehensivesearch of southern O star targets for rotational modulation.
To date 9 of the 11 confirmed magnetic O-type stars have been observed by
TESS .Of the sample HD 148937 ( P rot = 7.03 d), HD 47129 (1.21 d), HD 37742 (7.0 d) haverotational periods that are sufficiently short for them to have been observed for morethan a full rotational cycle (Petit et al., 2013; Grunhut et al., 2017). All 9 magneticstars show variability on timescales of 1-6 days and analysis is ongoing to determinethe source(s) of the variability. Here we discuss our preliminary findings from theanalysis of the magnetic O stars, which demonstrate the need for care in workingwith TESS photometry.Figure 1 shows the 2-min
PDCSAP light curve of CPD-59 2629 (Tr 16-22), amagnetic O8.5V star with a known rotation period of approximately 54 days (Naz´eet al., 2014). However, the principal variability in the extracted
TESS light curveappears to be due to a blended eclipsing binary. The peak of maximum power inL-S periodogram corresponds to a frequency of 0.87-d − . A pixel cutout of a sampleFFI image containing CPD-59 2629 (apparent g -band magnitude of 10.6) is overlayedwith all stars ( g <
11) in a 200 arcsec radius from a crossmatch with the Gaia DR2database using astroquery . A literature search shows that V731 Car ( g = 9 . − detected in the periodogram, and the light curve is shown phased to this period inthe bottom right panel of Figure 1. In fact, all stars shown on this FFI cutout are pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63) January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
11) in a 200 arcsec radius from a crossmatch with the Gaia DR2database using astroquery . A literature search shows that V731 Car ( g = 9 . − detected in the periodogram, and the light curve is shown phased to this period inthe bottom right panel of Figure 1. In fact, all stars shown on this FFI cutout are pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63) January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
Fig. 1:
Top : TESS light curve of CPD-59 2629 (Tr 16-22), sector 10, 2-min
PDCSAP flux.Variability does not appear to be from CPD-59 2629 but the nearby eclipsing binary V731Car.
Left : FFI pixel cutout centered on target. Stars shown are from crossmatch with GaiaDR2 in 200 arcsecond radius for apparent g -band magnitude less than 11. Numbers denote g -band magnitude and names come from crossmatch with SIMBAD using astroquery . Thecolour scale denotes flux. Top Right : L-S periodogram of light curve.
Bottom Right:
Light curve phased on P = 2.3 days, which corresponds to the known binary period ofV731 Car. Red circles denote binned points. In the analysis of HD 148937 (Fig. 2) we find a discrepancy between the 2-minprocessed light curve, and 30-min FFI extracted light curve obtained with eleanor .Figure 2 shows the 2-min light curve for HD 148937, a magnetic O6f?p type star witha 7.03-day rotation period (Wade et al., 2012). The L-S periodogram of the 2-mincadence data returns a period of P = 1 .
66 days, and the data are shown phasedto this period in the bottom left. Also shown is the 30-min eleanor
FFI extracted
CORR_FLUX light curve. It is generated from the raw FFI flux from an aperturecentered on the target and corrected for systematic effects. The L-S periodogramreturns a 6.97 day period, which is close to the published value of the rotationalperiod (7.03 d). The 2-min light curve is missing data from the start of the observingwindow and has a larger section removed in the middle. Other sources of discrepancy54 (cid:63)
PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 pta.edu.pl/proc/2020jan15/123
OBSTER: Identifying Candidate Magnetic O Stars through . . .
Fig. 2:
Top : TESS
Light curve of HD 148937 from 2-min cadence data, sector 12 (black).Below is the light curve extracted from 30-min FFI using eleanor (blue).
Left : Same asFig. 1 except for HD 148937.
Right : L-S periodograms of both light curves.
Bottom :Light curves phased to dominant periods, P = 1 .
66 d (2-min), P = 6 .
97 d (30-min). between the two methods may include the detrending of
TESS systematics, as wellas the difference in size and shape of the apertures. This shows that comparisons pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
Fig. 3: Gaia colour-magnitude diagram of southern 2-min
TESS targets crossmatched withGaia DR2 database using MAST (approx. 47 000 stars). Overlayed is the southern GOSCsample (315). Apparent g -band magnitudes are converted to absolute G -band magnitudesusing distances from Bailer-Jones et al. (2018). Spectral types other than O stars are colourcoded according to listed spectral type in SIMBAD. Our goal is to perform a comprehensive search for rotational modulation in Ostars in the southern ecliptic (sectors 1-13), utilizing both 2-min cadence processedlight curves and 30-min FFI extraction. The Galactic O-Star Catalog (GOSC; Ma´ızApell´aniz et al., 2013) provides spectral classifications of bright O-type stars andwill serve as the primary source for our sample. Other sample selection methodswere considered including crossmatching targets with SIMBAD, however, we havefound that SIMBAD may list old and inaccurate spectral classifications. This isespecially true for O stars, which often require high-resolution spectroscopy such asfrom the IACOB and OWN surveys (Sim´on-D´ıaz et al., 2015; Barb´a et al., 2017).Using Gaia DR2 photometry (Gaia Collaboration et al., 2018) we have placed oursample of southern GOSC stars on a colour-magnitude diagram of all southern 2-mincadence targets crossmatched from the Gaia DR2 database (Fig. 3). The apparent g -56 (cid:63) PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 pta.edu.pl/proc/2020jan15/123
OBSTER: Identifying Candidate Magnetic O Stars through . . .band magnitudes are converted to absolute magnitudes using distances determinedby Bailer-Jones et al. (2018) (approx. 47 000 stars). We note that we have notperformed any corrections for reddening or extinction. All targets are colour-codedaccording to their listed spectral type in SIMBAD except for the O-stars. Bluestars denote all southern GOSC targets that also have distance determinations fromBailer-Jones et al. (2018) (315 targets). It is apparent from these diagrams that itwould be difficult to perform a sample selection of O stars using a colour-magnitudeselection.
Fig. 4: Same as Fig. 1 for HD 199579. The frequency of max power corresponds to a periodof 4.68 days. The next 4 most significant peaks found from a pre-whitening procedurecorrespond to periods between 1.1 and 2.9 days.
Figure 4 shows an example target (HD 199579) that we have flagged as potentiallyshowing rotational modulation with a period of 4.68 days. This star had previouslybeen determined to be a candidate magnetic star by Grunhut et al. (2017), andmay warrant follow up spectropolarimetric observations. It also appears to be mul-tiperiodic and show stochastic low-frequency variability similar to other OB stars(Bowman et al., 2019a).
Further work is required to confirm the suitability of the 2-min processed
TESS light curves in the search for O star rotational modulation. Comparisons to light pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)
January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63) python tool we have developed to generate the pixel images seen in Figs. 1, 2and 4. We will perform a Lomb-Scargle analysis on all southern O star targets inour GOSC sample. A large number of GOSC stars have spectra available from theIACOB and OWN surveys, and have been analyzed to determine stellar parameters(e.g. Holgado et al., 2018). Light curves that are determined to show variabilitydue to rotational modulation can be modelled using ADM to determine magneticfield geometry and strength and used to judge the feasibility of detection in spec-tropolarimetric observations. Analysis of these observations and any magnetic fielddetections will provide valuable insight into the photometric signature of rotationallymodulated O stars and will refine our method for future studies.
Acknowledgements.
GAW acknowledges Discovery Grant support from the Natural Sciencesand Engineering Research Council (NSERC) of Canada. ADU gratefully acknowledges thesupport of the Natural Science and Engineering Research Council of Canada (NSERC). Theresearch leading to these results has received funding from the European Research Coun-cil (ERC) under the European Union’s Horizon 2020 research and innovation programme(grant agreement No. 670519: MAMSIE). S-SD acknowledges support from the SpanishGovernment Ministerio de Ciencia, Innovaci´on y Universidades through grant PGC-2018-091 3741-B-C22. This research includes data collected with the TESS mission, obtainedfrom the MAST data archive at the Space Telescope Science Institute (STScI). Fundingfor the TESS mission is provided by the NASA Explorer Program. STScI is operatedby the Association of Universities for Research in Astronomy, Inc., under NASA contractNAS 5–26555. This research has made use of the SIMBAD database, operated at CDS,Strasbourg, France.
References
Aerts, C., Rogers, T. M., ApJ , 2, L33 (2015)Albacete Colombo, J. F., Morrell, N. I., Niemela, V. S., Corcoran, M. F., MNRAS , 1,78 (2001)Alecian, G., Stift, M. J., A&A , A53 (2010)Bailer-Jones, C. A. L., et al., AJ , 2, 58 (2018)Barb´a, R. H., Gamen, R., Arias, J. I., Morrell, N. I., in J. J. Eldridge, J. C. Bray, L. A. S.McClelland, L. Xiao (eds.) The Lives and Death-Throes of Massive Stars,
IAU Sympo-sium , volume 329, 89–96 (2017)Blomme, R., et al., A&A , A4 (2011)Borra, E. F., Landstreet, J. D., Mestel, L., ARA&A , 191 (1982)Bowman, D. M., et al., A&A , A77 (2018)Bowman, D. M., et al., Nature Astronomy , 760 (2019a)Bowman, D. M., et al., A&A , A135 (2019b)Buysschaert, B., et al., MNRAS , 1, 89 (2015)Castro, N., et al., A&A , A81 (2015)David-Uraz, A., et al., arXiv e-prints arXiv:1912.01102 (2019a) (cid:63) PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 pta.edu.pl/proc/2020jan15/123
OBSTER: Identifying Candidate Magnetic O Stars through . . .
David-Uraz, A., et al., MNRAS , 1, 304 (2019b)David-Uraz, A., et al., arXiv e-prints arXiv:1912.02687 (2019c)Feinstein, A. D., et al., PASP , 1003, 094502 (2019)Gaia Collaboration, et al., A&A , A1 (2018)Grunhut, J. H., et al., MNRAS , 2, 2432 (2017)Holgado, G., et al., A&A , A65 (2018)Jenkins, J. M., et al., in G. Chiozzi, J. C. Guzman (eds.) Software and Cyberinfrastruc-ture for Astronomy IV, volume 9913, 1232 – 1251, International Society for Optics andPhotonics, SPIE (2016), URL https://doi.org/10.1117/12.2233418
Lecoanet, D., et al., ApJ , 1, L15 (2019)Ma´ız Apell´aniz, J., et al., in Massive Stars: From alpha to Omega, 198 (2013)Martins, F., et al., MNRAS , 3, 1423 (2010)Munoz, M. S., et al., MNRAS 2573 (2019)Naz´e, Y., Wade, G. A., Petit, V., A&A , A70 (2014)Owocki, S. P., et al., MNRAS , 4, 3830 (2016)Pedersen, M. G., et al., ApJ , 1, L9 (2019)Petit, V., et al., MNRAS , 1, 398 (2013)Ricker, G. R., et al.,
Journal of Astronomical Telescopes, Instruments, and Systems ,014003 (2015)Schneider, F. R. N., et al., Nature , 7777, 211 (2019)Shultz, M. E., et al., MNRAS , 3, 4154 (2019)Sikora, J., et al., MNRAS , 4, 4695 (2019)Sim´on-D´ıaz, S., et al., in Highlights of Spanish Astrophysics VIII, 576–581 (2015)Stassun, K. G., et al., AJ , 3, 102 (2018)Stibbs, D. W. N., MNRAS , 395 (1950)ud-Doula, A., Owocki, S. P., ApJ , 1, 413 (2002)Ud-Doula, A., Owocki, S. P., Townsend, R. H. D., MNRAS , 3, 1022 (2009)VanderPlas, J. T., ApJS , 1, 16 (2018)Villebrun, F., et al., A&A , A72 (2019)Wade, G. A., MiMeS Collaboration, Review: Magnetic Fields of O-Type Stars, Astronom-ical Society of the Pacific Conference Series , volume 494, 30 (2015)Wade, G. A., et al., MNRAS , 3, 2459 (2012)pta.edu.pl/proc/2020jan15/123 PTA Proceedings (cid:63)
January 15, 2020 (cid:63) vol. 123 (cid:63)(cid:63)