This thesis focuses on the implementation and application of Technology Computer-Aided Design (TCAD) numerical framework tasks for Ultra Large Scale Integration (ULSI) characterization. Using a standard formulation of a TCAD model, the diverse mathematical task algorithms are described with particular emphasis on their attributes within a TCAD environment. These include:
The software implementation issues of integrating task programs code within the VISTA framework are discussed. A client/server architecture is proposed. It is based on the Model Object entity, an object-oriented representation of a TCAD model. Accordingly, a uniform interface mechanism is built on the available framework infrastructure to supply model evaluation services for task clients. This facilitates the development of new task code as well as the integration of existing programs.
Various applications of characterization methodologies are presented. A new technique for the determination of the one- and two-dimensional MOSFET doping profiles via inverse modeling is described. The method fills an important gap in the available process metrology tools. TCAD simulators calibration is illustrated by three examples. A TCAD based worst case technology characterization procedure uses the task modules in order to predict and specify the extremes of the region of fluctuations of a given CMOS process. Finally, a polysilicon gate length determination technique which can be used to measure intra-die length variation is presented.
These applications highlight the importance of
TCAD tasks by supplying critical information for sub-micron CMOS
technology characterization and development.