Direct-search deep level photothermal spectroscopy: An enhanced reliability method for overlapped semiconductor defect state characterization

摘要

A method for resolving highly overlapped defects in rate-window analysis is proposed. This method offers high defect-state characterization reliability because it is based on direct multiparameter fitting of deep level photothermal spectra using combined temperature and frequency scans. Two direct search optimization algorithms are utilized as follows: the genetic algorithm for a search of possible solution areas and the pattern search algorithm for a refined search of global minimum. Four defect levels are identified using this technique.