Chemical vapor deposition
multiscale problem
approximation methods
numerical simulation
Model of PECVD apparatus: Verification and Simulations
Juergen Geiser
Geiser
Juergen
V. Buck
Buck
V.
M. Arab
Arab
M.
Institut für Mathematik, HumboldtUniversität zu Berlin (ISSN 08630976), 45 pp.
Model of PECVD apparatus: Verification and Simulations
Juergen Geiser
,
V. Buck,
M. Arab
Preprint series:
Institut für Mathematik, HumboldtUniversität zu Berlin (ISSN 08630976), 45 pp.
MSC 2000
 35K25 General theory of higherorder, parabolic equations

35K20 Boundary value problems for secondorder, parabolic equations
Abstract
In this paper we present the simulation
of a chemical vapor deposition for metallic bipolar plates.
For chemical vapor deposition, the delicate optimization between
temperature, pressure and plasma power is important to
obtain a homogeneous deposition, see \cite{hlava95}.
The aim is to reduce reallife experiments of a given CVD plasma reactor,
based on a large physical parameter space we have a hugh amount of
experiments.
A detail study of the physical experiments on a CVD plasma reactor
allows to reduce to an approximated mathematical model,
which is the underlying transportreaction model.
Significant region of the CVD apparatus are approximated and
physical parameters are transferred to the
mathematical parameters.
Such approximation reduced the mathematical parameter space to a
realistic amount of numerical experiments.
Based on interpolation and regression functions we fit
to the physical parameter space and can give first prediction
to deposition rates with the simulation model.
Here numerical experiments help to understand the deposition process and the
control the positions of the sources for the deposition and precursor gases.
For the simulations we apply analytical as well as numerical
methods to obtain results to predict the growth of thin
layers.
The results are discussed with physical experiments to
give a valid model for the assumed growth.
Here an important transfer of engineering research on modelling
reallife processes to achieve a simulatable mathematical model.
Such a model can be solved by numerical solvers and discretisation
schemes. The results can be used to obtain a new understanding of the
technical processes in engineering research.
This document is wellformed XML.