THE CASE AGAINST GRAPHITE PARTICLES IN INTERSTELLAR SPACE * (Letter to the Editor)

Graphite flakes are excluded as a major component of interstellar grains because their ultraviolet extinction peaks occur at wavelengths longward of 2700 A. Mie calculations for spherical graphite particles are shown to have no physical or astronomical relevance. Solid particles condensing in low-density mass flows from carbon stars are likely to be in the form of single crystals of graphite rather than polycrystalline soot as occurs most commonly in laboratory situations (Hoyle and Wickramasinghe, 1962; Donn et al., 1968). Soot particles condense under high density conditions such as in flames, whereas astrophysical condensation processes proceed under quiescent conditions that tend to favour the production of single crystals. Such crystals are mostly in the form of flakes, and less frequently in the form of hollow helical tubes (Ijima, 1991). In an earlier paper it was shown that graphite needles and hollow graphite tubes possess extinction properties that are, in general, inconsistent with astronomical data (Jazbi et al., 1991). Astronomical observations for a large number of stars show a more or less symmetrical hump of extinction centred at a wavelength close to 2175 A with a full width at half maximum of about 1.1 ~tm- 1. For cylindrical graphite tubes extinction profiles are significantly asymmetric, and furthermore peaks at a wavelength close to 2175 A are possible only with somewhat unrealistic constraints on size parameters. We are thus left to consider interstellar graphite either in the form of spheres or flakes. From 1965 onwards the attention of astronomers has been focussed on graphite spheres for the reason that rigorous extinction calculations are possible for this case (Wickramasinghe and Guillaume, 1965). It is, however, well known that graphite is strongly anisotropic in its optical and electrical properties. Graphite flakes are com- prised of planar linkages of hexagonal carbon rings, stacked in parallel planes. The electrical conductivity along the planes is high, similar to a metal, but in directions transverse to the planes the conductivity is about 100 times lower and the material behaves essentially like a dielectric. The best determinations of the complex refractive index of graphite are available for the case of light with electric vector parallel to the basal planes (Taft and Phillipp, 1965; see also Hoyle and Wickramasinghe, 1991). It has become customary to postulate the existence of graphite spheres with a complex dielectric constant equal to that determined by Taft and Phillipp holding for all directions of the electric vector of incident light. With