Application of nonlinear regression in the analysis of relaxation photocurrent waveforms

This article discusses the performance of an algorithm for detection of defect centers in semiconductor materials. It is based on direct parameter approximation with nonlinear regression to determine the parameters of thermal emission rate in the photocurrent waveforms. The methodology of the proposed algorithm was presented and its application procedure was described and the results of its application can be seen in measured photocurrent waveforms of a silicon crystal examined with High-Resolution Photoinduced Transient Spectroscopy (HRPITS). The performance of the presented algorithm was verified using simulated photocurrent waveforms without and with noise at the level of 10-2. This paper presents for the first time the application of the direct approximation method using modern regression and clustering algorithms for the study of defect centers in semiconductors.

---

This paper was founded by the grant no. UGB 737/2022. Additionally, the authors would like to thank the team of prof. Paweł Kamiński from the Łukasiewicz Institute of Microelectronic and Photonics for sharing the measurement results of the silicon Si sample tested with the HRPITS method.