Context. Strong, globally-organized magnetic fields are found for a small fraction of O, B, and A stars. At the same time, many theoretical and indirect observational studies have suggested the ubiquitous presence of weak localized magnetic fields at the surfaces of massive stars. However, no direct detections of such fields have been reported yet. Aims. We have carried out the first comprehensive theoretical investigation of the spectropolarimetric observational signatures of the structured magnetic fields. These calculations are applied to interpret null results of the recent magnetic surveys of massive stars. Methods. The intensity and circular polarization spectra of early-type stars were simulated using detailed polarized radiative transfer calculations with LTE model atmospheres. Similar to observational analyses, the mean Stokes I and V line profiles were obtained by applying a multiline averaging technique. Different spectropolarimetric observables were examined for multiple realizations of randomly distributed radial magnetic field spots on different spatial scales. Results. We characterize the amplitude of the circular polarization profiles and the mean longitudinal magnetic field as a function of magnetic spot sizes. The dependence of these observables on the effective temperature, projected rotational velocity, and inclination angle is also investigated. Using results of the recently completed Magnetism in Massive Stars (MiMeS) survey, we derive upper limits on the small-scale magnetic fields compatible with the MiMeS nondetections. Conclusions. According to our simulations, existing spectropolarimetric observations of sharp-lined massive stars rule out the presence of any small-scale fields stronger than 50-250 G, depending on the typical spot sizes. For broad-lined stars, the observations constrain such fields to being below approximately 1 kG.

Context. Most O-type stars and many B stars show unexplained cyclical variability in their spectral lines, i.e., modulation on the rotational timescale, but not strictly periodic. The variability occurs in the so-called discrete absorption components (DACs) that accelerate through the UV-wind line profiles and also in many optical lines. For such OB stars no dipolar magnetic fields have been detected with upper limits of ∼300 G. Aims. We investigate whether multiple magnetic loops on the surface rather than non-radial pulsations or a dipolar magnetic field can explain the observed cyclical UV and optical spectral line variability. Methods. We present time-resolved, high-resolution optical spectroscopy of the O6.5I(n)fp star λ Cephei. We apply a simplified phenomenological model in which multiple spherical blobs attached to the surface represent magnetic-loop structures, which we call stellar prominences, by analogy with solar prominences. We compare the calculated line profiles as a function of rotational phase, adopting a rotation period of 4.1 d, with observed relative changes in subsequent quotient spectra. Results. We identify many periodicities in spectral lines, almost none of which is stable over timescales from months to years. We show that the relative changes in various optical absorption and emission lines are often very similar. Our proposed model applied to the He ii λ4686 line can typically be fitted with 2−5 equatorial blobs with lifetimes between ∼1 and 24 h.Conclusions. Given the irregular timescales involved, we propose that the azimuthal distribution of DACs correspond to the locations of stellar prominences attached to the surface. This could explain the observed variability of optical and UV lines, and put constraints on the strength and lifetime of these structures, which can be compared with recent theoretical predictions, in which bright magnetic surface spots are formed by the action of the subsurface convection zone.

Many O and B stars show unexplained cyclical variability in their winds, i.e. modulation of absorption features on the rotational timescale, but not strictly periodic over longer timescales. For these stars no dipolar magnetic fields have been detected, with upper limits below 300 G. Similar cyclical variability is also found in many optical lines, which are formed at the base of the wind. We propose that these cyclical variations are caused by the presence of multiple, transient, short-lived, corotating magnetic loops, which we call "stellar prominences". We present a simplified model representing these prominences to explain the cyclical optical wind-line variability in the O supergiant λ Cephei. Other supporting evidence for such prominences comes from the recent discovery of photometric variability in a comparable O star, which was explained by the presence of multiple transient bright spots, presumably of magnetic origin as well.