From N~3 to N: pushing through the border between scalar and exact modeling methods

摘要

Formulating diffraction/scattering problems from the standpoint of general sources opens an algorithmic path where powerful ultra-fast numerical schemes can extensively be used for the exact modeling of unusually large sections of optical systems which have so far been only accessible by scalar methods. Breaking through the N~3 time and N2 memory barriers to a linear dependence will have an impact in DUV and EUV lithography, in possibly non-periodic large-NA DOEs as well as in light trapping/extracting scattering layers.