16.1: Introduction

With the exception of a few subjects in Chapter 7, we have dealt exclusively with spherical stars throughout this book. In addition, these stars have been "normal" stars; that is, they are stars without strange peculiarities or large time-dependent effects in their spectra. These are stars, for the most part, with interiors evolving on a nuclear time scale and atmospheres which can largely be said to be in LTE. The prototypical star that fits this description is a main sequence star, and for such stars the theoretical framework laid down in this book will serve quite well as a basis for understanding their structure and spectra. However, the notion of a spectral peculiarity is somewhat "in the eye of the beholder". Someone who looks closely at any star will find grounds for calling it peculiar or atypical. This is true of even the main sequence stars, and we would delude ourselves if we believed that all aspects of any star are completely understood.

Phenomena are known to exist in the atmospheres of stars that we have not included in the basic theory. For example, convection does exist in the upper layers of late-type stars and contributes to the transport of energy. However, the efficiency of this transport is vastly inferior to that which occurs in the stellar interior, because the mean free path for a convective element is of the order of that for a photon. Nevertheless, in the lower sections of the atmosphere, convection may carry a significant fraction of the total energy. Unfortunately, the crude mixing length theory of convective transport that was effective for the stellar interior is insufficient for an accurate description of atmospheric convection.

The turbulent motion exhibited by the atmospheres of many giant stars also represents the transport of energy and momentum in a manner similar to convection. Just the kinetic energy of the turbulent elements themselves should contribute significantly to the pressure equilibrium in the upper atmosphere. However, if one naively assigned a turbulent pressure of \(\rho\mathrm{v}^2\) to the elements, the measured turbulent velocities would imply a pressure greater than the total pressure of the gas and the atmosphere would be unstable. Clearly a better theory of turbulence is required before any significant progress can be made in this area.

In the last 20 years, it has become clear that the atmospheres of many stars are indeed unstable. Ultraviolet observations of the spectra of early-type stars suggest that they all possess a stellar wind of out-flowing matter of significant proportions. As one moves upward in the atmosphere of a star, departures from LTE increase to such a point that they dominate the atmospheric structure. In addition, energy sources such as magnetic fields, atmospheric mass motions, and possibly acoustic waves associated with convection or turbulence may contribute significantly to the higher level structure. While these energy contributions are tiny when compared to the energy densities in the photosphere, the contributions may supply the majority of the interaction energy between the star and its outermost layers which are basically transparent to photons.

In the hottest stars, the pressure of radiation, transmitted to the rarefied upper layers of the atmosphere through absorption by the resonance lines of metals, will cause this low-density region to become unstable, expand, and eventually be driven away. This is undoubtedly the source of the stellar winds for stars on the upper main sequence and in the early-type giants and supergiants. The extended atmospheres of these stars pose an entirely new set of problems for the theory of stellar atmospheres.

Many stars exist in close binary systems so that the atmosphere of each is exposed to the radiation from the other. In some instances, the illuminating radiant flux from the companion may exceed that of the star. Under these conditions we can expect the local atmospheric structure to be dominated by the external source of radiation. In addition, such stars are liable to be severely distorted from the spherical shape of a normal star. This implies that the defining parameters of the stellar atmosphere (\(\mu\), \(g\), and \(T_\mathrm{e}\)) will vary across the surface. Even in the absence of a companion, rapid rotation of the entire star can provide a similar distortion with the accompanying variation of atmospheric parameters over the surface.

These are some of the areas of contemporary research and to even recount the present state of progress in them would require another book. Instead of skimming them all, in this last chapter, we consider one or two of them in the hopes of pricking the reader's interest to find out more and perhaps contribute to the solution of some of these problems.