CPO: Milestone in the simulation of SEM images; a speedup in calculation time of 1600 times!
In nano lithography, SEM image simulators can be of great benefit in the study of dimensional metrology.An example is the interpretation of the true size, shape and roughness characteristics of three dimensional resist features in top-down SEM images. In particular, full Monte-Carlo simulation programs for SEM image acquisition are known to be notoriously slow.
Our quest in reducing the computation time of SEM image simulation further has led us to investigate the use of graphics processing units (GPUs) for metrology.
We have succeeded in creating a full Monte-Carlo electron-matter interaction simulator with sophisticated physical scattering models, which runs entirely on a GPU from NVIDIA.
As a case study for the performance, we simulated the exposure of a complex feature: an isolated silicon line with rough sidewalls located on a flat silicon substrate (see image).
The top-down SEM image is obtained with an Intel Xeon X 5650 processor in approx. 2 weeks.
That very same image is calculated in approx. 18 minutes using our newly developed simulator running on a GeForce GTX970 from NVIDIA.
This is a milestone in the simulation of SEM images: a speedup in calculation time of 1600 times!
We now have the potential to investigate case studies in CD-SEM metrology, which otherwise would take unreasonable amounts of computation time.
The new GPU simulator, including an application in nano lithography, will be presented at the SPIE conference for Advanced Lithography (21-25 February 2016) in San Jose, California.
drs. Thomas Verduin (T.Verduin@tudelft.nl)
Delft University of Technology
Faculty of Applied Sciences
Department of Imaging Physics
Charged Particle Optics Group
Lorentzweg 1, Room F127
2628 CJ Delft