Research On Infrared Thermal Wave Detection Of Micro-crack Defects In Semiconductor Silicon Wafers By Laser Excitation

As the main substrate material for integrated circuits,semiconductor silicon wafers have become the semiconductor materials with the largest production scale,the largest single crystal diameters,and the most complete production processes.Surface / sub-surface damage generated during the processing of semiconductor wafers will have a direct impact on the processing time and processing efficiency of subsequent polishing steps.Because the surface layer of the wafer is less damaged and has the crystallographic properties of single crystal,Many conventional inspection techniques are not suitable for wafer surface / subsurface damage inspection.Therefore,non-destructive testing,analysis,and evaluation of surface / sub-surface damage during the production of semiconductor wafers is necessary to achieve high-efficiency,and low-damage processing of silicon wafers.LFM infrared thermal wave nondestructive testing is an emerging nondestructive testing method.Compared with conventional detection technology,it has many advantages such as non-contact,large single-shot detection area.The laser has the characteristics of good monochromaticity,strong directivity,energy concentration and good coherence,and it is easy to obtain a uniform temperature field.The thesis mainly researches the detection technology of micro-crack defects in semiconductor silicon wafers by chirped laser-induced infrared thermal waves from the aspects of theoretical analysis,simulation research,testing experiments,and image processing.The heat transfer theory is used to analyze the heat generation and heat conduction process of micro-crack defects on semiconductor wafers.A three-dimensional heat conduction model is established and finite element simulation analysis is performed to obtain the temperature field distribution on the surface of semiconductor wafers.By processing the surface temperature signal,the influences and laws of the laser power parameters,modulation parameters and microcrack geometric parameters on the temperature difference were obtained.A semiconductor silicon wafer microcrack defect test piece was prepared,a chirped beam-splitting laser-excited infrared thermal wave detection system was built,which mainly included LFM laser excitation system,multi spot laser beam excitation optical system,infrared thermal imaging acquisition system and image sequence processing and analysis system,and a software for generating infrared thermal wave detection excitation system signals was developed based on the G language development environment,and a more systematic inspection test was carried out the study.Based on the background reduction of the test data,the Fourier transform,fractional Fourier transform,and principal component analysis algorithms were used to process the feature information of the surface of the silicon wafer.The signal-to-noise ratios processed by different algorithms are compared.The influences of laser power,modulation parameters and micro-crack geometric parameters on detection results are discussed.The edge detection hybrid algorithm based on gray level transformation-Iterative threshold-Canny operator for defect edge detection operator is researched.As a result,compared with the classic edge detection operator,the algorithm can more effectively implement micro-crack defects on semiconductor wafers' edge recognition,which lays a foundation for the subsequent quantitative identification of defects.