2011A&A...534A.127O

The purpose of the present investigation is to probe the dust properties inside a molecular cloud, in particular how particle growth and the presence of ice coatings may change the overall shape of the extinction curve. Field stars behind a molecular cloud can be used to probe the cloud extinction for both the reddening and the absorption features. By combining multi-colour photometry and IR spectroscopy the spectral class of the star can be determined as can the extinction curve, including the vibrational bands of ices and silicates. Based on observations of field stars behind the dark globule B335, we determine the reddening curve from 0.35 to 24µm. The water ice band at 3.1µm is weaker (τ(3.1)=0.4) than expected from the cloud extinction (A_V∼10 for the sightline to the most obscured star). On the other hand, the CO ice band at 4.7µm is strong (τ(4.67)=0.7) and indicates that the mass column density of frozen CO is about the same as that of water ice. We fit the observations to model calculations and find that the thin ice coatings on the silicate and carbon grains (assumed to be spherical) lower the optical extinction by a few percent. We show that the reddening curves for the two background stars, for which the silicate band has been measured, can be accurately modelled from the UV to 24µm. These models only include graphite and silicate grains (plus thin ice mantles for the most obscured star), so there is no need for any additional major grain component to explain the slow decline of the reddening curve beyond the K band. As expected, the dust model for the dense part of the cloud has more large grains than for the outer regions. We propose that the well established shallow reddening curve beyond the K band has two different explanations: larger graphite grains in dense regions and relatively small grains in the diffuse ISM, giving rise to substantially less extinction beyond the K band than previously thought. For the sight line towards the most obscured star, we derive the relation A_Ks=0.97.E(J-K_Ks), and assuming that all silicon is bound in silicates, N(2H₂+H)∼1.5x10²¹.A_V∼9x10²¹.A_Ks. For the rim of the cloud we get A_Ks=0.51.E(J-K_s), which is close to recent determinations for the diffuse ISM. The corresponding gas column density is N(2H₂+H)∼2.3x10²¹.A_V∼ 3x10²².A_Ks.