Research On The Influence Of Triangle Defect On The Characteristics And Reliability Of SiC JBS Diode

Compared with silicon materials,silicon carbide materials have many excellent characteristics,but also have higher defect density.These material defects,including structural defects and morphological defects,are still the key factors affecting the yield and reliability of SiC power devices.In the process of silicon carbide epitaxial growth,various structural defects in the silicon carbide substrate will first transform in the substrate epitaxial surface or directly extend into the epitaxial layer.When the structural defects grow to the epitaxial surface,they may transform into visible morphological defects,causing greater harm to the device characteristics.In this paper,the correlation of Carrot,Macro Pit and Triangle defect with failure and performance degradation of SiC JBS device is analyzed statistically.It is found that Triangle defect has the greatest influence on the device,followed by Macro Pit and Carrot defect is the smallest.Then,all kinds of morphological defects on SiC JBS were observed to determine the morphological characteristics of various defects,and the most fatal triangular defects were comprehensively classified: triangular defects with particles at the top(Class A),triangular defects with folds at the top(class B)and triangular defects with micro pits at the top(class C).Type A ? and type C ? triangular defects were studied for the first time.Using molten KOH etching,PL and Raman scattering techniques,it is proved that different kinds of triangular defects have different sources,which may come from BPD,TED,SF,TD and droppings,but they are all related to 3C-SiC.By using OBIRCH technology,it is proved that type A ? triangular defects are the most fatal,followed by type C ?,type A ? and type B triangular defects.These research results expand the source and influence of triangular defects,rank the lethality of different kinds of triangular defects,and have a more comprehensive and in-depth understanding of triangular defects in SiC materials.Next,the device with Triangle defect but not yet degraded was selected for 1000 hours of high temperature reverse bias experiment.It was found that even if the Triangle defect did not affect the device characteristics,more than 83% of the device reliability would still have hidden dangers,such as direct failure,gradual deterioration of leakage,large leakage level or no influence.These situations may occur in the reliable area It may also occur after the stress is stopped during the experiment.This experimental phenomenon is attributed to the difference in the location and types of triangular defects.When different kinds of triangular defects are located in different parts of the active region or the terminal region,the performance of SiC JBS diodes is different after long-term operation,and there will be varying degrees of degradation.Finally,the LOF method based on density outlier detection and KNN method considering the absolute difference between the data and its neighborhood are combined to predict the reliability risk of SiC JBS diode.The screening results further confirm the influence of triangular defects on the device.In order to avoid the error and contingency of the traditional LOF-KNN method,an evaluation method combining LOF-KNN algorithm and triangle defect location is proposed to screen out SiC JBS diodes with Triangle defects and abnormal LOF value.Because the Triangle defect probability will make the reliability of SiC JBS diode abnormal,compared with the LOF-KNN algorithm or Triangle defect location to determine whether the device has a reliability risk,this method has higher accuracy.