DEEP-LEVEL-TRANSIENT SPECTROSCOPY - SYSTEM EFFECTS AND DATA-ANALYSIS

Studies of typical DLTS systems using either the dual-channel boxcar averager or the lock-in amplifier techniques are presented. The effects of nonzero gate width for the boxcar averager, phase angle adjustment for the lock-in amplifier, and response time of a typical commercial capacitance meter are investigated. Errors introduced in the measurements by these effects are calculated for typical cases. Measurements of the gold donor level in silicon are presented, along with calculated corrections. We find the corrections to be minimal in the boxcar-averager method, but significant in the lock-in amplifier approach.