COMPARISON OF COVARIANCE LINEAR PREDICTIVE MODELING TO THE MODULATIONFUNCTION-METHOD FOR USE IN DEEP-LEVEL TRANSIENT SPECTROSCOPY

This article provides a quantitative understanding of the merits and d isadvantages of various methods of analyzing the deep level transient spectroscopy (DLTS) transients and spectrums. The modulation function method is compared to the covariance method of linear predictive model ing (CMLPM) transient analysis technique, and conventional DLTS spectr um methods. It is found that for high signal to noise ratio transients (high trap concentration), the modulation function method is superior to all other methods. As the signal to noise ratio decreases, CMLPM r esolves closely spaced defect levels better than any other method exam ined. For very low signal to noise ratios, the effects of noise can on ly be overcome by the inherent averaging present in the more conventio nal DLTS spectrum analyses. The proper choice of modulation functions is also discussed, and a modified modulation function is proposed for higher resolution of closely spaced traps. Finally, DLTS transients fr om an electron irradiated sample containing five traps, are analyzed b y conventional DLTS spectrum methods, the modulation function method a nd CMLPM to experimentally support the theoretical predictions.