We use a 3D Monte Carlo radiative transfer code to study the projection of large shadows by circumstellar disks around young stellar objects on surrounding reflection nebulosity .
It is shown that for a wide range of parameters a small ( 10-100 AU ) circumstellar disk can project a large ( 1 000-10 000 AU ) dark band in the near-infrared that often resembles a massive edge-on disk .
The disk shadows are divided into two basic types , depending on the distribution of the reflecting material and the resulting morphology of the shadows in the near-infrared .
Two YSOs associated with bipolar nebulosity , CK 3/EC 82 illuminating the Serpens Reflection Nebula ( SRN ) and Ced 110 IRS 4 in the Chamaeleon I molecular cloud , are modelled in detail as disk shadows .
Spectral energy distributions of the two sources are collected using both archival ISO data and new Spitzer-IRS data .
An axisymmetric model consisting of a small disk and a spherically symmetric envelope can reproduce the near-infrared images and full spectral energy distributions of the two disk shadow candidates .
It is shown that the model fits can be used to constrain the geometry of the central disks due to the magnifying effect of the projection .
The presence of a disk shadow may break a number of degeneracies encountered when fitting to the SED only .
Specifically , the inclination , flaring properties and extinction toward the central star may be independently determined from near-infrared images of disk shadows .
Constraints on the disk mass and size can be extracted from a simultaneous fit of SEDs and images .
We find that the CK 3 disk must have a very low mass in opacity-producing , small ( \lesssim 10 \mu m ) dust grains ( corresponding to a total mass of \sim 7 \times 10 ^ { -6 } { M _ { \odot } } , assuming a gas-to-dust ratio of 100 ) to simultaneously reproduce the very strong silicate emission features and the near-infrared edge-on morphology .
Ced 110 IRS 4 requires that a roughly spherical cavity with radius \sim 500 AU centered on the central star-disk system is carved out of the envelope to reproduce the near-infrared images .
We show that in some cases the bipolar nebulosity created by a disk shadow may resemble the effect of a physical bipolar cavity where none exists .
We find that a disk unresolved in near-infrared images , but casting a large disk shadow , can be modelled at a level of sophistication approaching that of an edge-on disk with resolved near-infrared images .
Selection criteria are given for distinguishing disk shadows from genuine large disks .
It is found that the most obvious observable difference between a disk shadow and a large optically thick disk is that the disk shadows have a compact near-infrared source near the center of the dark band .
High resolution imaging and/or polarimetry should reveal the compact source in the center of a disk shadow as an edge-on disk .
Finally , it is shown that disk shadows can be used to select edge-on disks suitable for observing ices located inside the disk .