A NEW FIGURE OF MERIT AND METHODOLOGY FOR QUANTITATIVELY DETERMINING DEFECT RESOLUTION CAPABILITIES IN DEEP-LEVEL TRANSIENT SPECTROSCOPY ANALYSIS

Citation
Wa. Doolittle et A. Rohatgi, A NEW FIGURE OF MERIT AND METHODOLOGY FOR QUANTITATIVELY DETERMINING DEFECT RESOLUTION CAPABILITIES IN DEEP-LEVEL TRANSIENT SPECTROSCOPY ANALYSIS, Journal of applied physics, 75(9), 1994, pp. 4570-4575
Citations number
32
Categorie Soggetti
Physics, Applied
Journal title
ISSN journal
00218979
Volume
75
Issue
9
Year of publication
1994
Pages
4570 - 4575
Database
ISI
SICI code
0021-8979(1994)75:9<4570:ANFOMA>2.0.ZU;2-9
Abstract
This article provides guidelines for accurately comparing the defect r esolution capabilities of different deep level transient spectroscopy (DLTS) analysis techniques. Additionally, guidelines for choosing meas urement parameters such as the proper weighting times and temperature range in order to achieve maximum energy resolution are presented. The modulation function method is compared to the correlation method of l inear predictive modeling (CMLPM) transient analysis technique, and co nventional DLTS spectrum methods. It is found that for extremely high signal to noise ratio transients (high trap concentration), the modula tion function method is superior to all other methods. As the signal t o noise ratio is reduced to a more practical range, CMLPM resolves clo sely spaced defect levels better than any other method examined. For v ery low signal to noise ratios, the effects of noise can only be overc ome by the inherent averaging present in the more conventional DLTS sp ectrum analyses. To demonstrate the applicability of this figure of me rit, a discussion of the resolution problems involved in distinguishin g between simultaneously present Ti and Mo impurities in Si is also pr esented. Finally, CMLPM is also found to be superior to the inverse la place transform technique.