A simulated annealing-simplex downhill hybrid algorithm is presented to solve the problems of ellipsometric data inversion. Basing on Monte Carlo technique of simulated annealing algorithm, the hybrid algorithm uses simplex downhill algorithm to get the local optimization and avoids the local optimization to get the global optimization by Metropolis accepting principle, then the global optimum ellipsometric data are obtained quickly. A typical model with single-layer absorbing film was dealt with by the hybrid algorithm and the simulated annealing algorithm respectively in numerical simulation experiments. The results show that the hybrid algorithm is feasible, credible and ascendant in ellipsometric data inversion. Furthermore, with the same testing conditions and inversion precision, the hybrid algorithm can save time with two quantity degrees, so it will be found more applications in practice.

Date Published: 20 November 2009
PDF: 6 pages
Proc. SPIE 7511, 2009 International Conference on Optical Instruments and Technology: Optoelectronic Measurement Technology and Systems, 75111E (20 November 2009); doi: 10.1117/12.837845

Published in SPIE Proceedings Vol. 7511:
2009 International Conference on Optical Instruments and Technology: Optoelectronic Measurement Technology and Systems
Shenghua Ye; Guangjun Zhang; Jun Ni, Editor(s)