Abstract details

For a wide interval, ranging from the far IR (1 mm) up to the far UV (100 \\A) , we plot the depth of formation of the individual wavelength points of a stellar spectrum onto a reference depth scale which may be either the column mass or the Rosseland mean optical depth. This is the by-product of our computation of stellar atmosphere models by means of a novel code that we developed with the aim of ascertaining the weight of the different physical phenomena upon the structure of a stellar atmosphere and its spectrum: a numerical laboratory for the study of stellar atmospheres. Such a code allows us to solve without any kind of difficulty the monochromatic radiative transfer (RT) equations within an interval of optical depths large enough to embrace the region of formation of the spectrum for each of the frequencies considered. Due to the enormous difference among the opacities at different frequencies (up to ten orders of magnitude between strong lines and optically thin continua), for any physical layer the monochromatic optical depth $\tau_{\nu}$ will vary accordingly. Therefore the condition $10^{- 3}\ \le\ \tau_{\nu}\ \le\ 10^{3}$ , that sets the limits for effective radiative transfer frequency by frequency, will hold at very different depths in the atmosphere. Consequently the reference scale (either a mean optical depth or the column mass) shall cover in most cases some 16 decades. Our code, due to the intrinsic nature of the numerical algorithm employed to solving the RT equations can cope with such demanding conditions, what the codes currently in use cannot. Thus we are in a position to study easily both radiative transfer and the energy balance inside stellar atmospheres. The first practical result of such an exercise is the possibility of quantifying the requirements for the correct computation of the emergent spectrum at wavelengths far from the visible region.