An observational evaluation of magnetic confinement in the winds of BA supergiants

Magnetic wind confinement has been proposed as one explanation for the complex wind structures of supergiant stars of spectral types B and A. Observational investigation of this hypothesis was undertaken using high-resolution (λ/Δλ ∼ 65 000) circular polarization (Stokes V) spectra of six late B- and early A-type supergiants (β Ori, B8Iae; 4 Lac, B9Iab; η Leo, A0Ib; HR1040, A0Ib; α Cyg, A2Iae; ν Cep, A2Iab), obtained with the instruments ESPaDOnS and Narval at the Canada–France–Hawaii Telescope and the Bernard Lyot Telescope. Least-squares deconvolution (LSD) analysis of the Stokes V spectra of all stars yields no evidence of a magnetic field, with best longitudinal field 1σ error bars ranging from ∼0.5 to ∼4.5 G for most stars. Spectrum synthesis analysis of the LSD profiles using Bayesian inference yields an upper limit with 95.4 per cent credibility on the polar strength of the (undetected) surface dipole fields of individual stars ranging from 3 to 30 G. These results strongly suggest that magnetic wind confinement due to organized dipolar magnetic fields is not the origin of the wind variability of BA supergiant stars. Upper limits for magnetic spots may also be inconsistent with magnetic wind confinement in the limit of large spot size and filling factor, depending on the adopted wind parameters. Therefore, if magnetic spots are responsible for the wind variability of BA supergiant stars, they likely occupy a small fraction of the photosphere.