Magnetism in Massive Stars


Science Summary

The MiMeS Project


  2. Magnetic fields are to astrophysics, as sex is to psychology.


  4. ⎯ H.C. van der Hulst, 1987


  6. Motivation


  8. Massive stars are those stars with initial masses above about 8 times that of the sun, eventually leading to catastrophic explosions in the form of supernovae. These represent the most massive and luminous stellar component of the Universe, and are the crucibles in which the lion’s share of the chemical elements are forged. These rapidly-evolving stars drive the chemistry, structure and evolution of galaxies, dominating the ecology of the Universe - not only as supernovae, but also during their entire lifetimes - with far-reaching consequences.


  1. In our sun and essentially all other cool, low-mass stars, vigorous magnetic activity results from the conversion of convective and rotational mechanical energy into magnetic energy, generating complex and variable magnetic fields whose properties correlate strongly with stellar mass, age and rotation rate. Although the “dynamo” mechanism that drives this process is not understood in detail, its basic principles are well established (e.g. Doble
    r 2005). The magnetic fields of hot, higher-mass stars are qualitatively different from those of cool, low-mass stars (e.g. Wade 2003). They are detected in only a small fraction of stars, and they are structurally much simpler, and frequently much stronger, than the fields of cool stars. Most remarkably, their characteristics show no clear correlation with basic stellar properties such as age, mass or rotation (e.g. Mathys et al. 1997, Kochukhov & Bagnulo 2006, Landstreet et al. 2007, Hubrig et al. 2007).  The weight of opinion holds that these puzzling characteristics reflect a fundamentally different field origin: that the observed fields are not generated by dynamos, but rather that they are fossil fields; i.e. remnants of field accumulated or generated during star formation (e.g. Mestel 1999, Moss 2001, Ferrario & Wickramasinghe 2006). This relic nature potentially provides us with a powerful and unique capability: to study how magnetic fields evolve throughout the various stages of stellar evolution, and to explore how they influence, and are influenced by, the structural changes that occur during all phases of stellar evolution: from stellar birth to stellar death.


  3. 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. 2008b).  The Magnetism in Massive Stars (MiMeS) project represents a comprehensive, multidisciplinary strategy by the foremost international researchers in the physics of hot, massive stars to address the “big questions” related to the complex and puzzling magnetism of massive stars.


  5. Hot, massive stars represent unique targets for the study of stellar magnetism. Their strong, radiatively-driven winds couple to magnetic fields, generating complex and dynamic magnetospheric structures (e.g. Babel & Montmerle 1997a, Donati et al. 2002). Recent models and simulations (e.g. ud-Doula et al. 2006; Townsend et al. 2005, 2007) show that magnetic confinement of stellar winds can explain wind variability and X-ray emission as observed in large numbers of OB stars (e.g. Kaper & Fullerton 1998, Stelzer et al. 2005). The interaction of the wind with the magnetic field modifies mass loss, and may enhance the shedding of rotational angular momentum via magnetic braking (e.g. Weber & Davis 1967). As the evolution of massive stars is particularly sensitive to rotation and mass loss (e.g. Chiosi & Maeder 1986; Maeder & Meynet 2000), the presence of even a relatively weak magnetic field can profoundly influence the evolution of massive stars and their feedback effects, such as mechanical energy deposition in the interstellar medium (ISM) and supernova explosions (e.g. Ekstrom et al. 2008).


  7. In early 2008, MiMeS was awarded "Large Program" status at the Canada-France-Hawaii Telescope (CFHT), where we have been allocated 640 hours of dedicated time with the ESPaDOnS spectropolarimeter from late 2008 through 2012. In 2010, similar status was awarded to MiMeS at the Bernard Lyot Telescope, where we have been awarded 590h of dedicated time with the Narval spectropolarimeter from 2010 through 2012. This commitment of the observatories, their staff, their resources and expertise, allocated as a result of extensive international expert peer review of many competing proposals, is being 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.


  9. Science Drivers


  11. The MiMeS project aims to address four general scientific foci of immediate and general interest:


  13. The origin and evolution physics of fossil magnetic fields: Significant work remains in order to fully explore and test the fossil field hypothesis, to understand the detailed physics leadin
    g from the microgauss fields in the ISM to the kilogauss fields often observed in main sequence stars, and to explore the potential contribution of dynamo-generated fields (e.g. Featherstone et al. 2007). Theoretical work shows that even a modest magnetic field in a collapsing interstellar cloud core can lead to a `magnetic braking catastrophe' (Allen, Li, & Shu 2003; Mellon & Li 2008). The efficiency of magnetic braking can prevent the formation of a Keplerian circumstellar disk, in contradiction to observations (e.g., Andrews & Williams 2005). The resolution must lie in a phase of intense magnetic flux loss in the near-protostellar environment, which remains unmodeled in numerical simulations. New 3D numerical magneto-hydrodynamic (MHD) simulations of fossil field evolution (e.g. Braithwaite & Nordlund 2006) have shown that quasi- stable field configurations can exist in young stars with characteristics similar to those observed in magnetic massive stars. Concurrent development of sophisticated field mapping techniques (e.g. Kochukhov et al. 2004, Donati et al. 2006a) is setting the stage to allow a detailed confrontation of these new models with observations. Moreover, basic observational data are urgently required to provide guidance to theory (e.g. Auriere et al. 2007) and to aid in interpreting observed stellar properties, including the incidence fraction of magnetic massive stars, the variation of the incidence as a function of stellar mass, environment and age, and the distribution of field strengths and geometries. 


  15. The physics of atmospheres, winds, envelopes and magnetospheres of hot stars: The supersonic, radiatively driven winds of massive stars couple with their rapid rotation and strong magnetic fields, generating
    complex and dynamic magnetospheric structures whose observational signatures span the complete electromagnetic spectrum. Early models of the wind-field interaction (e.g., Babel & Montmerle 1997ab, Donati et al. 2001, 2002) illustrated the potential power of magnetically-confined wind and wind shock paradigms for understanding the observational properties of magnetic massive stars. The seeds sown by these initial studies have subsequently flourished, both in the form of sophisticated multi-D numerical MHD models that can furnish detailed, quantitative predictions of the structure and dynamics of massive-star magnetospheres (ud-Doula & Owocki 2002; Townsend, Owocki & ud-Doula 2007; ud-Doula, Owocki & Townsend 2008), and as semi-analytical treatments that offer significant insight into the underlying physics (Townsend & Owocki 2005). Initial applications of these models to archival datasets have proven remarkably successful (e.g., Gagne et al. 2005; Townsend, Owocki & Groote 2005). As the theoretical models become progressively refined, there are growing demands for high-quality observational data -- optical, UV and X-ray spectroscopy, precision multiband photometry, radio measurements, and detailed magnetic field topologies and abundance maps --  for the models' validation, calibration and exploration.


  17. The rotation and rotational evolution of massive stars: The interaction of the stellar winds of hot, massive stars with the magnetic field modifies mass loss, and may enhance the shedding of angular momentum via magnetic braking (e.g. direct evidence presented by Mikulasek et al. 2008). Among the intermediate-mass magnetic stars, the influence of magnetic braking (likely during pre-main sequence evolution) is obvious – their rotational velocities are on average 3 times slower than non-magnetic single stars of similar mass (e.g.
    Abt & Morrell 1995). Although magnetic massive stars are likely subjected to weaker pre-main sequence braking and even spin-up (Stepien 2000, 2002), the opportunity exists for significant braking to occur on the main sequence via their strong winds. Consequently, in massive stars our picture of the impact of magnetic fields on rotational evolution is much less clear. Whereas many magnetic helium-strong stars rotate with velocities typical of non- magnetic B stars, some rotate significantly more slowly (e.g. HD 184927, with Prot = 9.5 days). A similar diversity is observed in pre-main sequence massive stars (e.g. Alecian et al. 2008a). On the other hand, slow rotation is common in many of the recently-discovered main sequence magnetic massive stars (e.g. β Cep, τ Sco, θ1 Ori C, HD 191612). While this may be partly due to an observational bias (magnetic detections are easier for slower rotators), null results obtained from very high signal-to- noise ESPaDOnS data for some intermediate and fast rotators argue that this effect is real. This leads to basic questions regarding the rotational evolution of massive stars, in particular if slow rotation is the result of the presence of the magnetic field or, instead, if rotation and its early evolution determines the initial growth and/or decay of a fossil field (e.g. Auriere et al. 2007). 


  19. The evolution of magnetic OB stars and origin of the magnetic fields of neutron stars and magnetars: The evolution of hot, massive stars may be strongly modified by the presence of a surface magnetic field (e.g. Maeder & Meynet 2005). To what extent can fields outside the star tell us about fossil or dynamo fields inside
    the star, which can influence internal differential rotation and circulation currents? A related issue is the origin of the strong (1011-1015 G) magnetic fields of neutron stars, which represent the ultimate endpoint of the evolution of many massive stars. Are these magnetic fields fossils (e.g. Ferrario & Wickramasinghe 2006), or are they generated during the violent processes accompanying their formation in supernovae (e.g. Heger et al. 2005)? Magnetic fields will be important in core-envelope coupling in the later stages of evolution, and therefore for the birth period distribution of neutron stars and all manner of supernova models, including pulsar “kicks”. A big uncertainly is whether magnetars are born slowly-rotating or rapidly-rotating ("millisecond magnetars", e.g. Bucciantini et al. 2007), and the relationship to “hypernovae” and gamma-ray bursts (e.g. Yoon et al. 2008). Whatever their origin, the field characteristics of their main sequence (e.g. Petit et al. 2008) and post-main sequence progenitors (e.g. Wolf-Rayet stars; de la Chevrotiere et al., in prep.) represent important constraints and input for neutron star field generation and stellar evolution models. 


  21. Structure of the Large Programs


  23. As was successfully argued to the Large Program selection committees, the primary pathway to understanding the fundamental issues outlined above are the MiMeS Large Observing Programs at the CFHT and TBL. These two-component programs will provide the basic observational data required for this investigation, allowing us to obtain fundamental statistical information about the magnetic properties of the overall population of hot, massive stars (the Survey Component), while simultaneously providing for detailed investigations of individual objects (the Targeted Component).


  25. Targeted component: The targeted component (TC) will provide data to map the magnetic fields and
    investigate the physical characteristics of a small sample of individual magnetic stars of great interest, at the highest level of sophistication possible. The ~25 TC targets (most to be observed with ESPaDOnS, the remainder to be observed with the Narval instrument at Pic du Midi observatory) have been selected to allow us to investigate a variety of physical phenomena, most of which will allow us to directly and quantitatively confront models. From a larger pool of candidates, these targets were selected to best display the physics described above.


  27. Survey component: The 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.


  29. 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. The SC includes main sequence and evolved OB stars, including Be stars, generally located in the Galactic field, and selected from the IUE catalogue. These targets were selected to
    span the entire range of spectral type of interest, with a sufficient number of stars to provide useful information about the variety and dispersion of field and other physical properties as a function of spectral type. Furthermore, the selection of these targets from the IUE catalogue ensures the existence of UV spectra. For all stars, IUE spectra (which have already been analyzed by MiMeS team members) will be used in conjunction with ESPaDOnS data to derived accurate fundamental parameters, including surface abundances, mass loss rates and wind terminal velocities, using state-of-the-art non-LTE wind models computed with TLUSTY (Hubeny & Lanz 1995), CMFGEN (Hillier and Miller 1998) and Fastwind (Puls et al. 2005). We have attempted to avoid significant biases in our SC target selection, apart from the necessary choice of stars with favourable properties for detection. In addition, we include in the SC sample OB star members of 6 open clusters, ranging in age from 6.5 to 45 Myr, and 9 bright Wolf-Rayet stars. In combination with existing data (e.g. Alecian et al. 2008, Petit et al. 2008) this subsample will allow us to investigate the evolution of magnetism in massive stars throughout their pre-main sequence, main sequence (from 1 to 45 Myr) and post main sequence lifetimes, and to study the influence of formation environment on fossil field characteristics.


  31. Each SC target will normally be observed twice in Stokes V during the LP. 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 properties and X-ray emission, wind properties, rotation, variability and surface chemistry diagnostics.


  33. Data acquisition, processing and archiving


  35. All MiMeS data acquired at the CFHT are currently automatically processed with the Libre-ESpRIT package, then downloaded to the Canadian Astronomy Data Centre (CADC). Reduced MiMeS data are available from the proprietary MiMeS Data Archive, hosted by the CADC. Post-processing of reduced MiMeS data is performed using the MiMeS Massive Stars Pipeline at the Royal Military College of Canada. MiMeS Narval data is automatically reduced at TBL and transmitted to the MiMeS dedicated server, located at the Paris-Meudon Observatory in Paris, France.


  37. MiMeS Workshops


  39. The international MiMeS collaboration currently meets twice annually (in principle) to review MiMeS results and strategy. The first MiMeS workshop occurred in May 2008 at RMC in Kingston, while the second MiMeS workshop was at the Observatoire de Paris in May 2009. The third MiMeS workshop was organised in Waimea, Hawaii in late 2009, and the fourth workshop was held in July 2010 in Armagh, Northern Ireland.


  41. MiMeS timeline


  43. The CFHT MiMeS Large Program has been allocated 640 hours over 9 semesters. It began in August 2008 (semester 2008B), and will ostensibly end in January 2013 (semester 2012B). The TBL MiMeS Large Program has been allocated 590 hours over 4 semesters. It began in August 2010, and will ostensibly end in January 2013.


  45. MiMeS leadership and administration


  1. The MiMeS project and Large Program are administered by the MiMeS Sterring Committee:


  1. MiMeS PI: Gregg Wade

  2. Front-line/phase 2 (FL/P2) coordinator: Jason Grunhut

  3. Data Archiving: David Bohlender

  4. Field OB stars: Jean-Claude Bouret

  5. Cluster OB stars: Véronique Petit

  6. Be stars: Coralie Neiner

  7. Herbig Be stars: Evelyne Alecian

  8. Wolf-Rayet stars: Nicole St. Louis

  9. Modeling/data interpretation: Oleg Kochukhov, Rich Townsend

  10. Theory: Stéphane Mathis


  1. MiMeS is part of the international MagICS collaboration, which is overseen by its own steering committee.