The Profile Interchange Format (PIF) [Duv88]
is used as the primary data format for
representing wafer data and for exchanging simulation data between
tools.
It allows a hierarchical boundary representation of geometries
and defines a set of distributed or concentrated attributes on
geometry segments and on grids to store material and dopant
distribution data.
Grids are built from elements of a variety types, including
triangular and rectangular elements [Fas94].
Dopant concentration data, e.g., are subdivided into interstitial
and active fractions, with the total (chemical) concentration
of a dopant species being the sum of the two.
Figure 4.5 gives a summary of the objects found in a
typical PIF wafer file. Arrows represent references from an object
to another one. If a PIF object is a collection of smaller entities,
an index specifies the entity addressed
.
Figure 4.5:
PIF objects in PIF wafer state file.
Figure 4.6 shows the navigation tool provided in VISTA/SFC to view and manipulate PIF data.
Figure 4.6:
GUI for viewing PIF data. The left part of the window provides
navigation through the hierarchical data structure, the right part
shows the data in ASCII representation.
PIF data are accessed by applications using the PIF
Application Interface (PAI) [FFS
91], one of VISTA's
core components.
The original PIF specification [Duv88] leaves ample room for
defining different data models. The PAI implementation enforces some
conventions on the structure of PIF objects while still allowing
ambiguities in the definition of geometrical objects.
To avoid any ambiguities, a wafer state convention has been
devised that imposes
a number of restrictions on acceptable data representations for
a wafer state model and greatly
enhances the practical usability of the PIF format with respect to
inter-tool data exchange. The most important restrictions are given in
Table 4.2; see also [Rie96].
Table 4.2:
Original PIF specification vs. VISTA/SFC PIF wafer state convention.
For accessing PIF data, tools are either linked with the PAI or use external wrappers to convert data to the PIF format (Figure 4.7). The second approach is often preferable as no changes of tool code are necessary and tool upgrading is simplified.
Figure 4.7: Data level integration schemes.