3.2 Point Response Interface Method

There exist very sophisticated methods and functions to analytically model the ion implantation process, but there is no universal set of analytical parameters valid for all implantation conditions, because the moments of the distribution functions mutually depend on the implantation angle, the dose, the energy and the target material composition in a very complex manner. Therefore the analytical models often have to be re-calibrated using experimental results for a process window (set of reasonable process parameter intervals) in question before they can be applied to a specific problem.

The idea of the point response interface method developed by Burenkov [14] is to provide calculated point response functions instead of analytically modeled distribution functions. A physically based Monte-Carlo ion implantation simulation (eg. Sec. 3.3) is performed to calculate the point response function as will be demonstrated in detail in Sec. 4.5.3.

The major difficulty of this method is to determine positions at the surface of the simulation domain for the calculation of point response functions. These surface points have to be representative for the complete simulation domain. Up to now there is no algorithm to automatically determine such representative points. Additionally the first implementation of the point response interface method in the analytical ion implantation simulator IMP3D [55] supports only the use of one representative point ( $\Rightarrow$ one point response). It is obviously very complicated to identify a single representative point for the whole simulation domain in case of non-planar surfaces.

Nevertheless the point response interface method is capable to accurately predict impurity concentrations in the active area of a transistor in the vicinity of a mask edge as will be demonstrated in Sec. 4.5.3.