Commonly the term "Wafer" denotes a circular disk that serves as base material in the semiconductor fabrication process. Hundreds of process steps are performed on a Wafer to create devices like transistors or diodes. These devices are arranged in functional units, so called dies, each comprising an integrated circuit (IC). Depending on the functional complexity of the IC, the used process technology, and the size of the Wafer disk, there are up to several thousand dies that are produced on one Wafer. This amounts to a gigantic number of individual devices. In TCAD simulations only a very small number of devices or even just a fraction of a single device is considered. Compared to a circuit or logic simulator a TCAD simulator operates at a very low level of abstraction. Therefore, in the TCAD field the term Wafer is used to denote the very small fraction of a Wafer disk that is used during a TCAD simulation.
A TCAD conform Wafer description contains the geometry (topography) of the device structure, and quantities as they are used by the simulator models (e.g. dopant concentrations, forces, stress, ...). Boundary information is used to identify parts of the surface of a simulation domain, and is necessary to define the geometrical region for boundary conditions (e.g. contacts of a device). A boundary information may also hold properties like the material type of a contact or quantities like interface charges between two regions.
On the other hand side the Wafer description must not contain any model parameters, boundary conditions for the underlying equation system, or any other information that is of use for a certain simulator only. The data model is required to be as independent of any specific requirements of a simulator as possible. For the same reason the Wafer description also lacks any kind of circuitry information or macro model description as they are used by circuit simulators (SPICE).
2003-03-27