The properties of polarized light are fully described by the four Stokes parameters I, V, Q, and U which, in turn, are determined by the atmosphere's thermal and magnetic structure. Thus a comparison of observed with computed Stokes parameters as function of wavelength (`Stokes profiles') allows to draw conclusions regarding the structure of the atmosphere.
Stokes V profiles of many photospheric lines arising from solar network and plage regions measured with low spatial resolution with the Fourier Transform Spectrometer (FTS) at NSO/Kitt Peak revealed systematic asymmetries between the areas and amplitudes of the two lobes of the Stokes V profiles (Stenflo et al. 1984; for an overview, see Solanki 1993). Such asymmetries arise if gradients of the line-of-sight components of both magnetic field and flow velocity are present (Illing et al. 1975). The absence of a significant wavelength shift of the Stokes V zero crossing in the FTS data (Solanki 1986) may be interpreted as due to spatially separated regions of magnetic field and flow along the line of sight. Such a configuration occurs at the periphery of a vertically oriented magnetic flux tube: the field lines fan out with height owing to the decreasing external gas pressure forming a `magnetic canopy', i.e., a layer of more or less inclined magnetic field on top of a field-free region. If the flux tube is surrounded by a convective downdraft, a vertical ray penetrating the canopy and the deeper layers encounters strong gradients where the field strength drops rather abruptly to zero and, assuming the velocity within the magnetic region to be insignificant, the velocity rises abruptly at the same position. Grossmann-Doerth et al. (1988) have shown that the Stokes V profile of such a ray is significantly asymmetric with no zero-crossing shift.
Stokes V profiles calculated on the basis of flux tube models with external down-flow agree with the FTS data as regards sign and magnitude of area asymmetry. However, oscillatory motions within the flux tube are additionally required in order to reproduce the observed ratio between area and amplitude asymmetries (Grossmann-Doerth et al. 1989; Solanki 1989). The presence of significant motions within the magnetized plasma has also been inferred from the broad wings of observed Stokes V profiles (Solanki 1986). Moreover, such motions normally occur in numerical MHD simulations of flux tubes (Steiner et al. 1998). These results have been corroborated by polarimetric observations with spatial resolution of about (Grossmann-Doerth et al. 1996; Martínez Pillet et al. 1997), which yielded Stokes V profiles with wide ranges of asymmetry and zero-crossing shift values. As a consequence, the absence of a Stokes V zero-crossing shift in the FTS data has to be explained as an averaging effect.
Recently, Sigwarth et al. (1999; see also Sigwarth 1999) obtained Stokes profiles with about 0.8" spatial resolution from a `quiet' region and from a region of emerging magnetic flux. The authors found a significant fraction of Stokes V profiles with such strong asymmetry that they consist of virtually one lobe only (the extreme case of `Type-1 Profiles' reported by Sanchez-Almeida et al. 1996). Is it possible that such profiles are formed by the `canopy mechanism' as described above or are additional effects required, like the superposition of profiles arising from mixed polarity in the resolution element (Rüedi et al. 1992)? Or is magnetic micro-structuring of the atmosphere a prerequisite to the formation of such `abnormal' Stokes profiles, as claimed by Sánchez Almeida & Lites (1999)? We have studied a number of simple models in an attempt to resolve this problem.
Sect. 2 summarizes the observational evidence for extremely asymmetric Stokes V profiles. In Sect. 3 we investigate the conditions for the formation of one-lobe profiles on the basis of calculated Stokes V profiles for various idealized configurations with discontinuities in magnetic field and velocity. We discuss the results in Sect. 4 and present our conclusions in Sect. 5.
© European Southern Observatory (ESO) 2000
Online publication: May 3, 2000