Size distribution of interstellar particles. III. Peculiar extinctions and normal infrared extinction

The effect of changing the upper and lower size limits of a distribution of bare graphite and silicate particles with n(a)αa/sup -q/ is investigated. Mathis, Rumpl, and Nordsieck showed that the normal extinction is matched very well by having the small-size cutoff, a/sub -/, roughly-equal0.005 or 0.01 μm, and the large size a/sub +/, about 0.25 μm, and q = 3.5 for both substances. We consider the progressively peculiar extinctions exhibited by the well-observed stars, sigma Sco, rho Oph, and theta1 Ori C, with values of R/sub v/[equivalentA/sub v//E(B--V)] of 3.4, 4.4, and 5.5 compared to the normal 3.1. Two (sigma Sco, rho Oph) are in a neutral dense cloud; theta1 Ori C is in the Orion Nebula. We find that sigma Sco has a normal graphite distribution but has had its small silicate particles removed, so that a/sub -/(sil)roughly-equal0.04 μm if q = 3.5, or q(sil) = 2.6 if the size limits are fixed. However, the upper size limit on silicates remains normal. In rho Oph, the graphite is still normal, but both a/sub -/(sil) and a/sub +/(sil) are increased, to about 0.04 μm and 0.4 or 0.5 μm, respectively, if q = 3.5, or q(sil)roughly-equal1.3 if the size limits are fixed. In theta1 Ori, the small limit on graphite has increased to about 0.04 μm, or q(gra)roughly-equal3, while the silicates are about like those in rho Oph. The calculated lambda2175 bump is broader than the observed, but normal foreground extinction probably contributes appreciably to the observed bump. The absolute amount of extinction per H atom for rho Oph is not explained. The column density of H is so large that systematic effects might be present. Very large graphite particles (a>3 μm) are required to ''hide'' the graphite without overly affecting the visual extinction, but a normal (small) graphite size distribution is required by the lambda2175 bump. We feel that it is unlikely that such a bimodal distribution exists