The infrared emission from dust surrounding newly formed O stars

The dust cocoon around a newly formed O6 zero-age main sequence (ZAMS) star still embedded in its natal molecular cloud has been modeled for a wide range of radial density distributions and grain properties. We deduce the following general conclusions from these investigations. 1. The only models that satisfy all the observational constraints (i.e., fit the observed spectrum from 1 mm to 1 mu m, are consistent with observed near-IR optical depths, and have small angular diameters) require thin dust shells with relatively large inner radii composed of grains with about one-half the graphite/silicate abundance ratio of the MRN-DL mixture for the diffuse interstellar medium. 2. Constant density models produce the best overall fit to the observations. Steep negative density gradient models (rho[r] proportional to r(n) where n <= - 1) produce too much warm dust causing the calculated spectrum to peak short-ward of the observed peak near 100 mu m. Steep positive density gradient models (n >= 1) produce too little warm dust causing the near infrared flux densities to be smaller than those observed. 3. The warm dust cocoons are large, typical radius similar to 10(18) cm for an O6 ZAMS star. Large dust shells are required to produce enough cool dust for the far-IR spectrum to peak near 100 mu m, as observed. A large dust-free cavity is required to avoid too much hot dust close to the star, which would produce too much near-infrared emission. The apparent angular diameters are small in comparison to the projected outer radii of the dust shells. 4. The dust cocoons are optically thick at wavelengths shortward of similar to 7 mu m. 5. The average dust temperatures drop very steeply just outside the inner boundary of the dust shell to less than 100 K in less than 0.1% of the outer shell radius. More than 99.9% of the volume of the shell contains dust cooler than 100 K; the average temperature implied by the peak of the observed far-IR spectrum is similar to 30 K. 6. The addition of water ice mantles to refractory grain cores produces a strong 3.07 mu m absorption feature and weaker 12 and 45 mu m absorption features. Ice mantles do not appear to significantly alter either the average dust temperature with radius or the optical depth with frequency (except at ice resonances). The observed 10 mu m feature is well reproduced by bare silicate grains alone, any additional absorption due to water ice at 12 mu m makes the calculated feature broader and deeper than observed. 7. Lowering the dust sublimation temperatures from 2000 to 1500 K for graphite and from 1500 to 1000 K for silicates made little difference to the emergent spectrum or to the optical depths. 8. All our models imply rather massive dust shells, total mass >= 5000 M-circle dot.