Partners and International Organizations
(English)
|
Fraunhofer Gesellschaft IIS/B, Erlangen (D), ST Microelectronics, Crolles (F), INRIA, Rocquencourt (F), TU Wien IME, Wien (A), ISE AG, Zürich (CH)
|
Abstract
(English)
|
Objectives
The project MAGIC_FEAT responds to the growing need for full three dimensional process simulation in industry. Due to the advanced miniaturisation, the behaviour of semiconductor devices can only be modelled if three dimensional effects are considered. On the physical modelling, point defect diffusion and the mechanical state of the wafer have to taken into account.
Before the project started the bottlenecks in process simulation were, on one side, the fact that meshing was not robust enough for arbitrary geometries, especially - for time-dependent structures - and on the other side, that different physical models were not well enough integrated into the simulation environment. The aim of the project is to overcome these problems.
The project partners are chosen to combine institutes and universities (ETHZ, FhG, INRIA, TUV) with expertise in process simulation and in meshing. An industrial partner (STM) and a software house (ISE AG) ensure that the final results will be commercial product.
The expected problem areas are:
· The need for anisotropic meshes at interfaces, pn-junctions, and diffusion fronts is in contrast with the Box-Method conforming
· Delaunay criterion demanded by the method employed to discretise of the partial differential equations.
· For time dependent problems like deposition, etching, and oxidation robust algorithms are needed to compute the wafer state at the new time steps.
· When re-meshing the structure, the interpolation error has to be minimised.
· The oxidation module must couple the mechanics equation with the dopant diffusion equations.
Approach
The software house (ISE AG) aims to integrate all physical models and meshing tools into their simulation environment.
The groups developing mesh generators have different approaches that focus on different properties of the mesh. They can coupled together so the strengths of each tool are used. The tool of ETHZ-IIS plays the role as the integration platform.
Work progress
The coupling of the different meshing engines into one tool is now in the evaluation phase. First results show that the tool is more robust than meshing algorithms available before the project. The convergence of the solver is also improved, which proves that mesh quality is sufficient.
Another breakthrough could be established, which is the use of third party solid modelling libraries to emulate etching and deposition. The created geometries are used to perform the benchmark computations.
The ion implantation module could be improved to incorporate more advanced function types.
Exploitation
The project finished its first feedback loop and could report successfully in the first review meeting. A public workshop was held with participants from European semiconductor companies. At ETHZ-IIS, two Ph.D. theses were published that are linked to the project.
|
