THE ORIGIN OF VARIATIONS IN THE 2175 ANGSTROM EXTINCTION BUMP

In this paper, small graphite grains are assumed to produce the 2175 Angstrom interstellar extinction bump. The principal observational characteristic of the bump in the invariance of its central wavelength (2174 +/- 9 Angstrom) among various lines of sight through diverse environments. On the other hand, its FWHM varies from about 0.80 mu m(-1) to 1.2 mu m(-1) among the same lines of sight. In this paper, the variations in bump width among different sight lines is attributed to coatings on the graphite grains. The necessary properties of the coating material are investigated. Calculations show that the mantle material must have a bulk resonance in the wavenumber range 4.4-5.0 mu m(-1) (approximate to 5.4-6.2 eV; lambda,approximate to 227-161 nm), and close to 4.8 mu m(-1) (lambda = 208 nm) if the resonance is strong. However, it need not be strong. Mantles with the required properties can explain: (a) the constancy of the x(0), the wavenumber of maximum extinction, for all sight lines except those with the broadest bumps; (b) the shift of x(0) to larger wavenumber for the broadest bumps; (c) the fact that the bumps in the regions of highest radiation density (bright nebulosities) are narrow, while those in dark, quiescent clouds are broad; (d) the lack of correlation of the bump width with R(v) (the total-to-selective extinction ratio), and, consequently, with the underlying UV extinction and particle size distribution; and (e) the behavior of the deviations of the bumps seen in individual sight lines from the mean. Mantles tend to broaden the bump, so bare grains are associated with the narrowest bumps (FWHM approximate to 0.80 mu m(-1)) The optical constants of Draine and Lee (1984) predict a bump for small bare graphite grains that is broader (FWHM approximate to 1.0 mu m(-1)) than the narrowest observed, with a central wavelength of 2100 Angstrom. Several alternate sets of optical constants (differing from the Draine-Lee values only in the neighborhood of the bump) are presented. All produce a bump with a FWHM = 0.80 mu m(-1) and a central wavelength of 2175 Angstrom. These constants, or some similar to them, must hold if the bump is produced from small bare graphite grains. Because of the uncertainties in the optical constants, the shapes of the graphite cores cannot be determined. However, the shapes must be very closely the same for all sight lines or there would be a strong variation in the central wavenumber of the bump. Simple neutral polycyclic aromatic hydrocarbons (PAHs) have measured laboratory absorptions remarkably similar to those required to explain the broadening of the bump. Only 1.5% of cosmic carbon in PAHs is required to produce very broad bumps. Alternatively, about 2%-4% of the combined abundances of the refractory elements Fe and Mg, more highly depleted onto grains in dense regions than in the diffuse interstellar medium, could explain the mantles if their absorption rises rapidly with energy, as it does for MgO. Coatings of amorphous carbon, hydrogenated or not, diamond-like bonded carbon, or water ice do not have published optical constants that vary rapidly enough in wavenumber to produce the variations.