Subsections

• Conclusion
• Outlook

11. Conclusion and Outlook

Conclusion

State of the art algorithms for two- and three-dimensional surface evolution processes, namely, deposition and etching processes were developed and implemented. The simulator is based on the level set algorithms and can be used to simulate all common deposition and etching processes. The level set method has been shown to be the best alternative for overcoming the problems emerging in other surface tracking methods, e.g., high memory consumption and CPU time. This improvement was obtained by using different techniques such as calculating the signed distance function by a fast marching algorithm, extending the speed function, and dynamically moving the narrow band according to the new zero level set.

A basis for understanding the desired and undesired effects on topography simulation during the semiconductor manufacturing processes was presented. The techniques for overcoming the undesired problems were discussed and were used for obtaining an etching profile with minmal corner rounding, for instance.

In order to obtain the best model for the deposition processes, different physical models were tested by using SIESTA which uses the inverse modeling technique. This helped to model two important processes, namely, the deposition of the silicon dioxide from TEOS and silicon nitride. These deposition processes were used to simulate the capacitances which contribute to the timing delays in interconnect lines. The final capacitance calculation was done by joining ELSA and RCX tools. The significant influence of void formation on the capacitances was quantified, as using voids in a controlled and reproducible manner can be an economically advantageous substitute for low-$k$ materials.

Furthermore the optimized parameters of the deposition models obtained by SIESTA were used in the three-dimensional topography simulator to predict the void chracteristics during interconnect processes. These characteristics enabled to set layout design rules depending on geometrical parameters to avoid the formation of cracks.

Finally a new level set based method to generate structurally aligned grids while guaranteeing quality criteria of the triangulation was presented. It provides a lot of flexibility, since the resolution and anisotropy of the grid is customizable and the diameter of the triangles may vary over several orders of magnitude within one simulation domain.

Outlook

Simulations of the deposition and etching processes allow to gain insight into the considered processes. The more chemistry is considered in physical models the more accurate is this insight. Therefore, physical models which are more related to the chemistry of the processes must be developed for the simulation of different deposition and etching processes. Simultaneously a balance between the complexity of the models and the simulation time has to be considered to avoid very long simulation times. The models should be first implemented in two dimensions because of simplicity. After successfully finding the models which are in a good agreement with measurements, they have to be implemented for three dimensional simulations.

Having a full three-dimensional process tool capable of the simulation of the whole process steps is becoming more and more important. Topography simulation is a very essential issue in the simulation of most of process steps. Therefore, the topography simulator has to be integrated into a fully three-dimensional process tool. In addition, the different physical models proposed and implemented in the future in ELSA will also be incorporated into a fully three-dimensional process tool.

Finally, the two- and three-dimensional development and implementation of a multi level set algorithm are planed. This is, for instance, absolutely essential for the simulation of plasma etching processes where different regions of materials and masks have to be defined. Furthermore, a multi level set appraoch has to be used for the simulation of growing a number of crystals. Each individual crystal grows until it hits another crystal, forming a grain boundary. The detection of a grain boundary and its later moving need to be handeld with multi level set method.