For years I have been modeling the data produced during semiconductor manufacturing. My data models have evolved and stabilized into the following model that I have named the "Semiconductor Data Object Model (SDOM)". The following is a overview of my complete model that will provide an introduction to the concepts of the SDOM.
The data produced during semiconductor manufacturing is identified by a combination of abstract and physical identifiers. The SDOM is a collection of object models for each of these identifiers. Additionally, there are object models that implement the data produced at each hierarchial identifier.
The following diagram is a representation of the the objects that have physical analogs and their related data objects. (The items enclosed in brackets ("[]") are data objects.)
DataSource [Attributes] Lot [Attributes] [Statistics] [Data] Wafer [CoordSystems] [Attributes] [Statistics] [Data] Site [Statistics] [Data] Reticle [Statistics] [Data] Die [Data] Part [Data] Part [Data] [ParamSpecs] [CondSpecs] [CoordSystems]
The primary objects are "data source" (DataSource), "parameter specifications" (ParmSpecs), "condition specifications" (CondSpecs), and "coordinate systems" (CoordSystesms). One DataSource object has attribute data and references to one or more "lot" (Lot) objects. Each Lot object has attribute, statistic, and value data and references to one or more "wafer" (Wafer) or "part" (Part) objects. Each Wafer object has coordinate system, attribute, statistic, and value data and references to one or more "site" (Site), "reticle" (Reticle), "die" (Die), and "part" (Part) objects. Each Site object has statistic and value data. Each Reticle object has statistic and value data. Each Die object has value data. Each Part object has value data.
The object relationships are depicted another way in the following relationship diagrams:
A more detailed view of the object relationships is depicted in the following class diagram:
The SDOM has proven extremely useful in the development of applications and databases as the fundamental data model for semiconductor manufacturing and engineering data.