|
3D Resist profile aware OPC has becoming increasingly important to address hot spots generated at etch processes
due to the mass occurrence of non-ideal resist profile in 28nm technology node and beyond. It is therefore critical to
build compact models capable of 3D simulation for OPC applications. A straightforward and simple approach is to
build individual 2D models at different image depths either based on actual wafer measurement data or virtual
simulation data from rigorous lithography simulators. Individual models at interested heights can be used by
downstream OPC/LRC tools to account for 3D resist profile effects. However, the relevant image depths need be
predetermined due to the discontinuous nature of the methodology itself. Furthermore, the physical commonality
among the individual 2D models may deviate from each other as well during the separate calibration processes. To
overcome the drawbacks, efforts are made in this paper to compute the whole bulk image using Hopkins equation in
one shot. The bulk image is then used to build 3D resist models. This approach also opens the feasibility of
including resist interface effects (for example, top or bottom out-diffusion), which are important to resist profile
formation, into a compact 3D resist model. The interface effects calculations are merged into the bulk image
Hopkins equation. Simulation experiments are conducted to demonstrate that resist profile heavily rely on interface
conditions. Our experimental results show that those interface effects can be accurately simulated with reference to
rigorous simulation results. In modeling reality, such a 3D resist model can be calibrated with data from discrete
image planes but can be used at arbitrary interpolated planes. One obvious advantage of this 3D resist model
approach is that the 3D model is more physically represented by a common set of resist parameters (in contrast to
the individual model approach) for 3D resist profile simulation. A full model calibration test is conducted on a
virtual lithography process. It is demonstrated that 3D resist profile of the process can be precisely captured by this
method. It is shown that the resist model can be carried to a different lithography process with same resist setup but
a different illumination source without model any accuracy degradation. In an additional test, the model is used to
demonstrate the capability of resist 3D profile correction by ILT.
|
