Dust extinction and absorption: the challenge of porous grains

¹ Sobolev Astronomical Institute, St. Petersburg University, Universitetskii prosp. 28, St. Petersburg, 198504 Russia e-mail: nvv@astro.spbu.ru
² Isaac Newton Institute of Chile, St. Petersburg Branch, Russia
³ Max-Planck-Institut für Astronomie, Königstuhl 17, 69117 Heidelberg, Germany

Received: 5 May 2005
Accepted: 9 August 2005

Abstract

In many models of dusty objects in space the grains are assumed to be composite or fluffy. However, the computation of the optical properties of such particles is still a very difficult problem. We analyze how the increase of grain porosity influences basic features of cosmic dust – interstellar extinction, dust temperature, infrared bands and millimeter opacity.
It is found that an increase of porosity leads to an increase of extinction cross sections at some wavelengths and a decrease at others depending on the grain model. However, this behaviour is sufficient to reproduce the extinction curve in the direction of the star σ Sco using current solar abundances. In the case of the star ζ Oph our model requires larger amounts of carbon and iron in the dust-phase than is available. Porous grains can reproduce the flat extinction across the wavelength range measured for several lines of sight by ISO and Spitzer.
Porous grains are generally cooler than compact grains. At the same time, the temperature of very porous grains becomes slightly larger in the case of the EMT-Mie calculations in comparison with the results found from the layered-sphere model. The layered-sphere model predicts a broadening of infrared bands and a shift of the peak position to larger wavelengths as porosity grows. In the case of the EMT-Mie model variations of the feature profile are less significant. It is also shown that the millimeter mass absorption coefficients grow as porosity increases with a faster growth occurring for particles with Rayleigh/non-Rayleigh inclusions. As a result, for very porous particles the coefficients given by two models can differ by a factor of about 3.