Multithread structure as a possible solution for the Lβ problem in solar prominences

Abstract

Following the pioneering works of Heasley, Mihalas, Milkey and Poland (see e.g. Heasley and Milkey, 1983) who built non LTE onedimensional models of solar prominence, much attention has been paid to the spectral signatures of the Lyman lines as observed with OSO 8 (Vial, 1982_a). In spite of a better treatment of the frequency redistribution and boundary conditions, one-dimensional low-pressure models lead to Lyman β intensities much lower than observed ones (Heinzel, Gouttebroze and Vial, 1987). Different atomic processes of formation of hydrogen lines (Cooper, Ballagh and Hubeny, 1988) or the inclusion of a Prominence Corona Transition Region or PCTR (Heinzel, Gouttebroze and Vial, 1988) have been proposed to explain this discrepancy. We present here a different approach where the filamentary nature of prominences which provides the hydrogen lines with different opacities, offers their photons different escaping possibilities. The thread models we use derive from an energy equation where radiative losses are balanced by conductive flux (Foutenla and Rovira, 1983, 1985). We show that no superposition of threads gives good values of Lyman a, β and H a intensities for too high and too low pressures. Solutions are found for pressure around 0.05–0.1 dyn/cm² and a number of threads between 100 and 400. Two improvements have been performed: first, the inclusion of Partial Redistribution leads to a decrease of Lα (and Lβ) intensity and models now require a higher number of threads; second, the inclusion of the ambipolar diffusion along the steep temperature gradient which changes the hydrogen ionization in the lower regions (Foutenla, Avrett and Loeser, 1990). The new run of temperature and density implies more material at low temperatures and hydrogen lines intensities increase. A solution for the Lβ problem can be found for a pressure of about 0.1 dyn cm^-t2. However the Hα intensity appears to be rather high. Moreover, the number of threads required (about 200) is far larger than the number derived by Zirker and Koutchmy (this issue) and Mein (this issue) from observed Hα profiles. Our neglect of the radiative interaction between threads may explain our results (Heinzel, this issue). To conclude, these computations of non-lte radiative transfer in realistic geometrical and physical models, appear to be a promising path for the investigation of solar prominences.