Goal
Provide optimization based tools to improve the
design and performance
of chemical vapor deposition (CVD) reactors
used in semiconductor processing.
Background and Project Description
Chemical vapor deposition (CVD) is used extensively in the
manufacturing of semiconductors.
As CVD reactors, such as
small-batch fast-ramp low pressure CVD furnaces
become increasingly expensive, there is a higher
payoff for increased reactor throughput or for an increase
in the number of usable wafers.
Computational simulation combined with optimization are powerful
and efficient tools to achieve these objectives.
To accurately describe the physics of the process,
increasingly complex mathematical models are used, which
lead to large and computing intensive simulation codes.
The implementation of these simulation codes alone is a challenging
task that can often only be performed with state-of-the-art HPCC technology.
A combination of such simulation codes with optimization algorithms
to produce usable tools for the designer or operator of CVD reactors
requires new and innovative ways in the formulation of optimization
algorithms, the use of HPCC technology,
and their interaction with various simulation codes.
The goal of this project is to provide an optimization
environment for the designer or operator
- which is modular and can be easily interfaced with
various simulation codes,
- which can be quickly adapted to varying constrained and unconstrained
optimization problems, and
- which can be used in a variety of computing environments.