«The MiMeS Project: Magnetism in Massive Stars G.A. Wade1, E. Alecian1,2, D.A. Bohlender3, J.-C. Bouret4, J.H. Grunhut1, H. Henrichs5, C. ...»
Cosmic Magnetic Fields: From Planets, to Stars and Galaxies
c 2009 International Astronomical Union
Proceedings IAU Symposium No. 259, 2009
K.G. Strassmeier, A.G. Kosovichev & J.E. Beckman, eds.
The MiMeS Project:
Magnetism in Massive Stars
G.A. Wade1, E. Alecian1,2, D.A. Bohlender3, J.-C. Bouret4, J.H.
Grunhut1, H. Henrichs5, C. Neiner6, V. Petit7, N. St. Louis8, M.
Auri`re9, O. Kochukhov10, J. Silvester1, A. ud-Doula11 e and the MiMeS Collaboration† 1 Royal Military College of Canada, 2 LESIA, France, 3 Canadian Astronomy Data Centre, LAM, France,5 Ast. Inst. Amsterdam, Netherlands, 6 GEPI, France, 7 Universit´ Laval, 4 e arXiv:0812.4078v1 [astro-ph] 22 Dec 2008 Canada, 8 Univ. de Montr´al, Canada, 9 LAT, France, 10 Uppsala University, Sweden, e 11 Morrisville State College, USA Abstract. The Magnetism in Massive Stars (MiMeS) Project is a consensus collaboration among the foremost international researchers of the physics of hot, massive stars, with the basic aim of understanding the origin, evolution and impact of magnetic ﬁelds in these objects. The cor- nerstone of the project is the MiMeS Large Program at the Canada-France-Hawaii Telescope, which represents a dedication of 640 hours of telescope time from 2008-2012. The MiMeS Large Program will exploit the unique capabilities of the ESPaDOnS spectropolarimeter to obtain
† www.physics.queensu.ca/∼wade/mimes 119 120 G.A. Wade et al.
structural changes that occur during all phases of stellar evolution, from stellar birth to stellar death.
Although this fossil paradigm provides a powerful framework for interpreting the magnetic characteristics of higher-mass stars, its physical details are only just beginning to be elaborated (e.g. Braithwaite & Nordlund 2006, Auriere et al. 2007, Alecian et al. 2008a).
In particular, our knowledge of the basic statistical properties of massive star magnetic ﬁelds is seriously incomplete. There is a troubling deﬁcit in our understanding of the scope of the inﬂuence of ﬁelds on massive star evolution, and almost no empirical basis for how ﬁelds modify mass loss.
The Magnetism in Massive Stars (MiMeS) Project represents a comprehensive, multidisciplinary strategy by an international team of recognized researchers to address the big questions related to the complex and puzzling magnetism of massive stars. Recently, MiMeS was awarded ”Large Program” status by both Canada and France at the CanadaFrance-Hawaii Telescope (CFHT), where the Project has been allocated 640 hours of dedicated time with the ESPaDOnS spectropolarimeter from late 2008 through 2012.
This commitment of the observatory, its staﬀ, its resources and expertise, allocated as a result of an extensive international expert peer review of many competing proposals, will be used to acquire an immense database of sensitive measurements of the optical spectra and magnetic ﬁelds of massive stars, which will be applied to constrain models of the origins of their magnetism, the structure, dynamics and emission properties of their magnetospheres, and the inﬂuence of magnetic ﬁelds on stellar mass loss and rotation and ultimately the evolution of massive stars. More speciﬁcally, the scientiﬁc objectives
of the MiMeS Project are:
• To identify and model the physical processes responsible for the generation of magnetic ﬁelds in massive stars;
• To observe and model the detailed interaction between magnetic ﬁelds and massive star winds;
• To investigate the role of the magnetic ﬁeld in modifying the rotational properties of massive stars;
• To investigate the impact of magnetic ﬁelds on massive star evolution,and the connection between magnetic ﬁelds of non- degenerate massive stars and those of neutron stars and magnetars, with consequential constraints on stellar evolution, supernova astrophysics and gamma-ray bursts.
2. Structure of the Large Program To address these general problems, we have devised a two-component Large Program (LP) that will allow us to obtain basic statistical information about the magnetic properties of the overall population of hot, massive stars (the Survey Component), while simultaneously providing detailed information about the magnetic ﬁelds and related physics of individual objects (the Targeted Component).
Targeted component: The MIMeS Targeted Component (TC) will provide data to map the magnetic ﬁelds and investigate the physical characteristics of a small sample of known magnetic stars of great interest, at the highest level of sophistication possible. The roughly 20 TC targets have been selected to allow us to investigate a variety of physical phenomena, and to allow us to directly and quantitatively confront models.
Each TC target will be observed many times using the ESPaDOnS spectropolarimeter, in order to obtain a high-precision and high-resolution sampling of the rotationallymodulated circular and linear polarisation line proﬁles. Using state-of-the-art tomographic reconstruction techniques such as Magnetic Doppler Imaging (Kochukhov & 121 The MiMeS Project Figure 1. Least-Squares Deconvolved proﬁles of 3 hot stars: θ1 Ori C (O7V, left), Par 1772 (B2V, middle) and NU Ori (B0.5V, right). The curves show the mean Stokes I proﬁles (bottom panel), the mean Stokes V proﬁles (top panel) and the N diagnostic null proﬁles (middle panel), black for 2006 January and red for 2007 March. Each star exhibits a clear magnetic signature in Stokes V. These results are representative of those expected from the MiMeS Survey Component.
From Petit et al. (2008).
Piskunov 2002), detailed maps of the vector magnetic ﬁeld on and above the surface of the star will be constructed.
Survey component: The MiMeS Survey Component (SC) will provide critical missing information about ﬁeld incidence and statistical ﬁeld properties for a much larger sample of massive stars. It will also serve to provide a broader physical context for interpretation of the results of the Targeted Component. From a much larger list of potential OB stars compiled from published catalogues, we have generated an SC sample of about 150 targets which cover the full range of spectral types from B2-O4 which are selected to be best-suited to ﬁeld detection. Our target list includes pre-main sequence Herbig Be stars, ﬁeld and cluster OB stars, Be stars, and Wolf-Rayet stars.
Each SC target will be observed once or twice during the Project, at very high precision in circular polarisation. From the SC data we will measure the bulk incidence of magnetic massive stars, estimate the variation of incidence versus mass, derive the statistical properties (intensity and geometry) of the magnetic ﬁelds of massive stars, estimate the dependence of incidence on age and environment, and derive the general statistical relationships between magnetic ﬁeld characteristics and X-ray emission, wind properties, rotation, variability, binarity and surface chemistry diagnostics.
Of the 640 hours allocated to the MiMeS LP, 385 hours are committed to the SC and 255 hours are committed to the TC. The TC commitment includes 50 hours reserved for follow-up of targets detected in the Survey Component.
3. Precision magnetometry of massive stars For all targets we will exploit the longitudinal Zeeman eﬀect in metal and helium lines to detect and measure photospheric or pseudo-photospheric magnetic ﬁelds. Splitting of a spectral line due to a longitudinal magnetic ﬁeld into oppositely-polarized σ components produces a variation of circular polarisation across the line (commonly referred to as a (Stokes V ) Zeeman signature or magnetic signature; see Fig. 1.). The amplitude and morphology of the Zeeman signature encode information about the strength and structure of the global magnetic ﬁeld. For some TC targets, we will also exploit the transverse Zeeman eﬀect to constrain the detailed local structure of the ﬁeld. Splitting of a spectral line by a transverse magnetic ﬁeld into oppositely-polarized π and σ components produces a variation of linear polarisation (characterized by the Stokes Q and U parameters) across the line (e.g. Kochukhov et al. 2004).
122 G.A. Wade et al.
Figure 2. Magnetic Doppler Imaging (MDI) of the B9p star HD 112413 (Kochukhov et al.
, in preparation), illustrating the reconstructed magnetic ﬁeld orientation (lower images) and intensity (upper images) of this star at 5 rotation phases. The maps were obtained from a time-series of 21 Stokes IQU V spectral sequences. Although the ﬁeld line orientation of HD 112413 is approximately dipolar, the ﬁeld intensity map is far more complex. Maps similar to these will be constructed for the MiMeS Targeted Component.
3.1. Survey Component For the SC targets, the detection of magnetic ﬁeld is diagnosed using the Stokes V detection criterion described by Donati et al. (1992, 1997), and the surface ﬁeld constraint characterised using the powerful Bayesian estimation technique of Petit et al. (2008).
After reduction of the polarized spectra using the Libre-Esprit optimal extraction code, we employ the Least-Squares Deconvolution (LSD; Donati et al. 1997) multi-line analysis procedure to combine the Stokes V Zeeman signatures from many spectral lines into a single high-S/N mean proﬁle (see Fig. 1), enhancing our ability to detect subtle magnetic signatures. Least-Squares Deconvolution of a spectrum requires a line mask to describe the positions, relative strengths and magnetic sensitivities of the lines predicted to occur in the stellar spectrum. The line mask characteristics are sensitive to the parameters describing the stellar atmosphere. In our analysis we employ custom line masks carefully tailored to best reproduce the observated stellar spectrum, in order to maximize the S/N gain of the LSD procedure and therefore our sensitivity to weak magnetic ﬁelds.
The exposure duration required to detect a Zeeman signature of a given strength varies as a function of stellar apparent magnitude, spectral type and projected rotational velocity. This results in a large range of detection sensitivities for our targets. The SC exposure times are based on an empirical exposure time relation derived from real ESPaDOnS observations of OB stars, and takes into account detection sensitivity gains resulting from LSD and velocity binning, and sensitivity losses from line broadening due to rapid rotation. Exposure times for our SC targets correspond to the time required to deﬁnitely detect (with a false alarm probability below 10−5 ) the Stokes V Zeeman signature produced by a surface dipole magnetic ﬁeld with a speciﬁed polar intensity. Although our calculated exposure times correspond to deﬁnite detections of a dipole magnetic ﬁeld, our observations are also sensitive to the presence of substantially more complex ﬁeld toplogies.
123 The MiMeS Project Figure 3. Example of the spectral and spatial emission properties of a rotating massive star magnetosphere modeled using Rigid Field Hydrodynamics (Townsend et al. 2007). The stellar rotation axis (vertical arrow) is oblique to the magnetic axis (inclined arrow), leading to a complex potential ﬁeld produced by radiative acceleration, Lorentz forces and centripetal acceleration. The consequent heated plasma distribution in the stellar magnetosphere (illustrated in colour/grey scale) shows broadband emission, and is highly structured both spatially and spectrally. Magnetically-conﬁned winds such as this are responsible for the X-ray emission and variability properties of some OB stars, and models such as this will be constructed for the MiMeS Targeted Component.
3.2. Targeted Component Zeeman signatures will be detected repeatedly in all spectra of TC targets. The spectropolarimetric timeseries will be interpreted using several magnetic ﬁeld modeling codes at our disposal. For those stars for which Stokes V LSD proﬁles will be the primary model basis, the modeling codes employed by Donati et al. (2006) or Alecian et al. (2008b) will be employed. For those stars for which the signal-to-noise ratio in individual spectral lines is suﬃcient to model the polarisation spectrum directly, we will employ the Invers10 Magnetic Doppler Imaging code to simultaneously model the magnetic ﬁeld, surface abundance structures and pulsation velocity ﬁeld (Piskunov & Kochukhov 2002, Kochukhov et al. 2004). The resultant magnetic ﬁeld models will be compared directly with the predictions of fossil and dynamo models (e.g. Braithwaite 2006, 2007, Mullan & Macdonald 2005, Arlt 2008).
Diagnostics of the wind and magnetosphere (e.g. optical emission lines and their linear polarisation, UV line proﬁles, X-ray photometry and spectroscopy, radio ﬂux variations, etc.) will be modeled using both the semi-analytic Rigidly-Rotating Magnetosphere approach, the Rigid-Field Hydrodynamics (Townsend et al. 2007) approach and full MHD simulations using the 3D ZEUS code (e.g. Stone & Norman 1992; see Fig. 3).
4. MiMeS data pipeline Following their acquisition in Queued Service Observing mode at the CFHT, ESPaDOnS polarised spectra are immediately reduced by CFHT staﬀ using the LibreEsprit reduction package and downloaded to the dedicated MiMeS Data Archive at the Canadian Astronomy Data Centre in Victoria, Canada. Reduced spectra are carefully normalized to the continuum using existing software tailored to hot stellar spectra. Each reduced ESPaDOnS spectrum is then subject to an immediate quick-look analysis to verify nominal resolving power, polarimetric performance and S/N. Preliminary LSD proﬁles are extracted using our database of generic hot star line masks to perform an initial magnetic ﬁeld diagnosis and further quality assurance. Finally, each ESPaDOnS spectrum will be processed by the MiMeS Massive Stars Pipeline (MSP; currently in production) to determine a variety of critical physical data for each observed target, in 124 G.A. Wade et al.
addition to the precision magnetic ﬁeld diagnosis: eﬀective temperature, surface gravity, mass, radius, age, variability characteristics, projected rotational velocity, radial velocity and binarity, and mass loss rate. These meta-data, in addition to the reduced high-quality spectra, will be uploaded for publication to the MagIcS Legacy Database†.