Abstract
Optimizing the properties of optical and photonic devices calls for the need to control and manipulate light within
structures of different length scales, ranging from sub-wavelength to macroscopic dimensions. Working at different
length scales, however, requires different simulation approaches, which have to account properly for various effects such
as polarization, interference, or diffraction: at dimensions much larger than the wavelength of light common ray-tracing
techniques are conveniently employed, while in the (sub-)wavelength regime more sophisticated approaches, like the socalled
finite-difference time-domain (FDTD) technique, are used.
Describing light propagation both in the (sub-)wavelength regime as well as on macroscopic length scales can only be
achieved by bridging between these two approaches. Unfortunately, there are no well-defined criteria for a switching
from one method to the other, and the development of appropriate selection criteria is a major issue to avoid a
summation of errors. Moreover, since the output parameters of one simulation method provide the input parameters for
the other one, they have to be chosen carefully to ensure mathematical and physical consistency.
In this contribution we present an approach to combine classical ray-tracing with FDTD simulations. This enables a joint
simulation of both, the macro- and the microscale which refer either to the incoherent or the coherent effects,
respectively. By means of an example containing one diffractive optical element (DOE) and macroscopic elements we
will show the basic principles of this approach and the simulation criteria. In order to prove the physical correctness of
our simulation approach, the simulation results will be compared with real measurements of the simulated device. In
addition, we will discuss the creation of models in FDTD based on different analyze techniques to determine the
dimensions of the DOE, as well as the impact of deviations between these different FDTD models on the simulation
results.
