Detection statistics of deep levels in minority carrier transient spectroscopy

A theoretical treatment of the minority carrier transient spectroscopy (MCTS) experiment is presented. We have modeled the minority carrier flux through the depletion region of an illuminated Schottky diode held under reverse bias, and used these data to calculate the occupancy of minority carrier traps as a function of energy, capture cross section, and temperature. The model shows that the capacitance transient monitored in the MCTS experiment decreases in intensity as the temperature is raised. It is demonstrated that this causes inaccuracies in the measured deep level activation energy E(a) derived from an Arrhenius plot of the data. Simulated MCTS spectra have been compared with measured MCTS spectra of hole emission from the gold donor in silicon, and very good agreement between modeled and experimental spectra is observed. The model explains the commonly observed phenomenon of a reduction in MCTS peak heights for increasing temperature, which is the opposite effect from that commonly observed in conventional deep level transient spectroscopy data. It is shown that the correct choice of rate window can significantly reduce errors in the measured value of E(a). (C) 1997 American Institute of Physics.