When coating a 3-dimensional object, certain areas of the surface may move into the shadows cast by protruding features of the object itself. Here, we speak of "self-shadowing". To correctly predict the thickness distribution we must take into account the blockage of the material vapor due to the self-shadowing.
In this example. a deep parabolic reflector is mounted on a planetary-rotation fixture in a magnetron sputtering chamber. As the reflector revolves in its orbit the bottom part of the bowl spends much of the time in the shadow cast by the reflector wall itself. To simulate the self-shadowing, we introduce a negative and rotational mask with a circular opening that matches the rim of the reflector. This shadow mask cast the same shadow as the wall of the reflector.
Because of the self-shadowing, the thickness in the center of the bowl is thinner compared to the wall of the parabola (X-Z map). Disregarding the self-shadowing by removing the shadow mask, the program yields a thickness distribution which is thicker in the center, in error. The thickness maps with and without self-shadowing are shown below.