Modern optical components require sophisticated coatings with tough specifications and the design of optical multilayers has become a key activity of most laboratories and factories. A synthesis technique based on the simulated annealing algorithm is presented here. In this stochastic minimization, no starting solution is required, only the materials and technological constraints need to be specified. Moreover, the algorithm will always reach the final result. As simulated annealing is a stochastic algorithm, a great amount of state transitions is needed in order to reach a global minimum of the merit function used to evaluate the difference between the optical target and the calculated filter. Anyway the computing time remains reasonable on a work-station. A few examples will show the performances of our program. It also has to be pointed out that no refinement is needed at the end of the annealing because the solution is already highly optimized. Nowadays the design of robust filters with low sensitivity to technological variations remains a key factor for manufacturers. This is why we have established some criteria that quantify the robustness of the stacks. It also enables comparison of multilayers synthesized by different methods and corresponding to the same target.

Date Published: 19 August 1996
PDF: 12 pages
Proc. SPIE 2776, Developments in Optical Component Coatings, (19 August 1996); doi: 10.1117/12.246838

Published in SPIE Proceedings Vol. 2776:
Developments in Optical Component Coatings
Ian Reid, Editor(s)