Comparison and evaluation of measured and simulated high-frequency capacitance-voltage curves of MOS structures for different interface passivation parameters

• Autor: Sevillano Bendezú, Miguel Ángel
• Formato: tesis de maestría
• Fecha de Publicación: 2019
• Institución: Pontificia Universidad Católica del Perú
• Repositorio: PUCP-Institucional
• Lenguaje: español
• OAI Identifier: oai:repositorio.pucp.edu.pe:20.500.14657/168082
• Enlace del recurso: http://hdl.handle.net/20.500.12404/14462
• Nivel de acceso: acceso abierto
• Materia: Semiconductores de óxido metálico; Semiconductores; Circuitos integrados--Pasivación; https://purl.org/pe-repo/ocde/ford#1.03.00

Semiconductor-insulator interfaces play an important role in the performance of many different electronic and optoelectronic devices such as transistors, LEDs, lasers and solar cells. Particularly, the recombination of photo-generated charge carriers at interfaces in crystalline silicon solar cells causes a dramatic efficiency reduction. Therefore, during the fabrication process, the crystalline silicon must be subjected to prior superficial passivation; typically through an insulating layer such as SiO2, SiNx or AlOx. The function of this passivating layer is to reduce electrical recombination losses in interfacial defect states originating from dangling bonds. The associated passivation parameters are, on the one hand, stable charges within the insulating layer (Qox) that by repelling a certain type of charge carrier from the crystalline silicon surface, reduces its recombination effectiveness (Field Effect Passivation). On the other hand, the density of surface defect states or the interface trap density (Dit), which is reduced by the passivation layer (Chemical Passivation). These passivation parameters (Qox and Dit) turn out to be relevant when evaluating the effectiveness of a new material with passivating properties, as well as relevant for different theoretical models that allow simulations of the spectral response and/or efficiency in solar cells under different passivation conditions. One of the techniques widely used for studying the interfacial passivation properties of semiconductor electronic devices is the extraction of these interfacial passivation parameters through of capacitance-voltage (C-V) measurements on metal-oxide-semiconductor (MOS) or metal-insulator-semiconductor (MIS) systems. In the present work, a simulation tool for High-Frequency C-V curves based on simulated Qox and the Dit was developed using Python. As a first step, the simulation was developed for an ideal MOS system, i.e. for Qox = 0 and Dit = 0. A verification of the resulting, simulated band-bending was reached through a band diagram simulator (The Multi-Dielectric Band-Diagram program). As a second step, the program was subjected to an evaluation and validation through experimental data. This data comprises measurements of C-V and their respective extracted parameters for a sample of silicon dioxide thermally grown on crystalline silicon wafer (SiO2/c-Si). Using three different models for the Dit distribution within the band gap energy: Gaussian model, U-shape model, and a constant value, approximations of the corresponding experimental C-V curve were obtained. It was evident that the C-V curve simulated from the Dit based on the model with Gaussian distributions for the defect centers and exponentials for the band tails resulted in the best approximation of the experimental C-V curve. It should be noted that the other two models were adjusted based on the value of the Dit near to midgap energy, where the recombination probability and rate are the highest. In this way, the constant model of the Dit at the midgap presented the largest deviation in the simulated C-V curve among the used models. An implicit fitting method of the Dit through the experimental C-V curve fitting is proposed. For this, the U-shape model is used because it only depends on three parameters. The average values of the fitted and the experimentally extracted Dit are compared. The parameter D0 it, which defines the value at midgap in the U-shape model could be interpreted as an average estimation of the Dit energetic range values around the midgap where recombinations are most significant. Therefore, this parameter could determine a representative value of the Dit. Finally, the developed program allows an in-depth analysis of the passivation parameters from which the surface passivation is evaluated.